Friday, July 09, 2010

Gender Imbalance in Science

There is a gender imbalance in science. This is most obvious when looking at math, physics or engineering programs. It's easy to look around biomedical labs, see the number of women, and assume this isn't a problem for the biological sciences. And indeed, in the biological sciences, women make up 62% of bachelor's recipients and 56% of graduate students1. Beyond that, the numbers drop off for a number of reasons. Again, in the biological sciences, only 41% of postdoctoral fellows, 33% of tenure-track and 22% of tenured scientists are women. (Nevermind the fact that their salaries also tend to be lower)

Those numbers are pretty shameful. Some might argue that it doesn't matter, as long as the science being done is good (though I'm really not sure who they might be). The problem is, it isn't. Even the research being done is gender biased, as highlighted by a series of recent editorials in Nature. The first points out that there's a gender bias in clinical trials: except for a few key areas (eg. breast, cervical or ovarian cancers) women are underrepresented
Generally, women remain underrepresented in biomedical research. Studies published in 2000 and in 2008 concluded that women were still not included in mixed-sex cardiovascular trials in numbers that reflect the disease prevalence among the general population. A survey of studies published in 2004 in nine influential medical journals found that only 37% of participants were women (24% when restricted to drug trials), and only 13% of studies analysed data by sex.
It's no secret that women and men have different physiologies, and can be differently affected by diseases and their treatments. So why should sex differences not be considered when designing a clinical trial? You can't treat differently if you don't know the differences, and even where gender is known to affect outcomes, there's still no difference in treatment
A 2005 study of 300 new drug applications between 1995 and 2000 found that even those drugs that showed substantial differences in how they were absorbed, metabolized and excreted by men and women had no sex-specific dosage recommendations on their labels. This may be part of the reason why women are 1.5 times more likely to develop an adverse reaction to prescription drugs than men.
And if you're a pregnant woman, it gets even worse. You're forced to either forego treatment or potentially settle for drugs of unknown effectiveness or safety (for you or the fetus), since pregnant women are usually excluded from trials and new drugs are often not approved for them. This despite the obvious fact, that the second article points out, that "pregnant women get sick, and sick women get pregnant." Obviously diving right into clinical trials with pregnant women is a bit risky, in no small part for the fetus (though a controlled clinical trial is probably still safer than rolling the dice with off-label use), but there are ways this can be approached.
There is an obvious alternative: small, well-designed trials for pregnant women, starting with phase I safety trials that would begin at the same time as phase III efficacy trials in the general population. With this staggered approach, pregnant women and fetuses would not be exposed to any compounds that failed in phase I and II trials.
This strategy is simple, and makes sense: it allows pregnant women the same standard of evidence-based care afforded to everybody else (even better, if you consider the lack of sex consideration described above), accounting for their own particular changes while mitigating risk.

Most surprising, to me at least, is the fact that gender bias even extends to animal models. In some ways, it makes sense to keep models simple and avoid potential complications from hormone cycles. Depsite the fact that male mice tend to fight when caged together and single housing takes both space and money, a survey of 2000 studies done in 2009 found a male bias in animals used.
We found a male bias in 8 out of 10 biological disciplines, most pronounced in neuroscience (5.5 males to 1 female), pharmacology (5 males to 1 female) and physiology (3.7 males to 1 female).
Though this might be more convenient or have a practical rationale, it doesn't necessarily result in the best model. So we have biased models feeding into imbalanced clinical trials, with women getting the short end at each step. And the end result:
However justifiable these imbalances may be on a case-by-case basis, their cumulative effect is pernicious: medicine as it is currently applied to women is less evidence-based than that being applied to men.
Which is just sad.

--
1All numbers based on US statistics from the NSF, 2006.


15 comments:

n/a said...

"The first points out that there's a gender bias in clinical trials: except for a few key areas (eg. breast, cervical or ovarian cancers) women are underrepresented"

Yeah, just a few minor exceptions like that:

Prostate Cancer incidence rates are comparable to Breast Cancer rates. Yet, federal funding is significantly higher for Breast Cancer research than for Prostate Cancer.

In 2008, there were the same number of new cases of Prostate Cancer (186,320) as Breast Cancer (184,450), as estimated by the American Cancer Society.

* In 2007, Breast Cancer received almost two thirds more funding ($127.5 million) from the Congressionally Directed Medical Research Program as did Prostate Cancer ($80 million dollars) (59.5% difference).

* In 2007, Breast Cancer received more than twice as much funding ($707 million) as did Prostate Cancer ($305 million) from our nation’s medical research agency, the National Institutes of Health (231% difference).


On the broader subject of "discrimination" against women in science, even The New York Times has more sense than you:

I’m all in favor of women fulfilling their potential in science, but I feel compelled, at the risk of being shipped off to one of these workshops, to ask a couple of questions:

1) Would it be safe during the “interactive discussions” for someone to mention the new evidence supporting Dr. Summers’s controversial hypothesis about differences in the sexes’ aptitude for math and science?

2) How could these workshops reconcile the “existence of gender bias” with careful studies that show that female scientists fare as well as, if not better than, their male counterparts in receiving academic promotions and research grants?

Kamel said...

n/a: I'm not talking about trial funding here, I'm talking about how they're set up. Do you really think it's unreasonable that for a disease that affects both genders, women are underrepresented in trials? Or, if you don't want to put it in terms of females, how is a trial that doesn't look at gender difference good for *either* sex? The funding of different research programs is a different question, but yes one worth thinking about.

Re: your second point: well, it's the NYT that isn't making sense. That quote, as a counterargument to what I've written, is saying that women are worse at science and math (poor aptitude) while simultaneously excelling at science and research (more promotions/grants). Which is it?

Besides, it misses the point. The barriers to women in science have nothing to do with aptitude or ability to get grants. It has everything to do with the million little ways the system, as it stands, works against them, from lack of mentors to worse pay to pressures to stay home and take on the domestic role, etc., etc., etc.

Look at the numbers I provided. Do you really think that all these women are able to get their advanced degrees but suddenly suck at science when it's time for a postdoc or beyond? The fact that women are as good or better at getting funding and promotions, yet are woefully underrepresented at the top levels of academia doesn't make you wonder why?

n/a said...

Kamel,

I suggest you watch this talk from a man who formerly sat on the board of directors of NOW.

Yes, a smaller proportion of women than men have the aptitude to do math and science at the highest levels.

If you have evidence women are paid less for the same work and level of experience, please present it.

I'm sure most women (excluding gender feminists) would find your presumption that women want (or should want) the exact same things as men strange, if not offensive. Many women prefer "domestic roles". Many don't want to devote their lives as exclusively to work as men do. Considering the overproduction of PhDs and the generally shitty career prospects in science, it might be more reasonable to argue that it's the men who disproportionately continue to postdoc and beyond who are getting shafted.

I'm all in favor of research taking into account sex differences where appropriate, but I'm not sure feminist graduate students are the best ones to be lecturing scientists on the topic.

MalePhD said...

"Yes, a smaller proportion of women than men have the aptitude to do math and science at the highest levels. " should probably read as: a smaller proportion of women then men [had]...

62% of bachelor's graduates are women, how is that a small proportion. The times they are a changing. The disparity of men and women entering bachelor's and graduate's levels is increasing, towards women. Arguments are being made that the teaching styles in high school are more beneficial to women, which would obviously have more of an effect then some innate differences in biology. In education nurture outweighs nature.

That's my proper response. My laughable response to n/a depends on whether he just implied Kamel is a feminist graduate student or not. If n/a did then hahaha. Literally nothing in that statement is true. Time for the big reveal Kamel... he's a man (shocking). Technically a scientist should have a perfect 50/50 split and this should be clearly stated in the methods, that is how biological science should be conducted. Arguments against (which are unfounded) should not come from people eager to label those who prefer proper scientific methods as feminists.

n/a's assumption relates too perfectly with the old riddle of a man getting in a car accident with his son, being sent to different hospitals and the son's surgeon saying 'I can't operate, he's my son'. How's that possible???

Side note: good post Kamel, I appreciate you not getting into implications of who is better at science or 'innate abilities' in the original post.

rob said...

Good post. I assume that even i/a agrees that the clinical and preclinical trial sex biases are inexcusable. [the main point of the post] I have to at least partially agree with him that women may make different career/lifestyle choices than men leading to sex differences in the workplace. However, these lifestyle choices are likely heavily influenced by societies traditional parental roles. In Sweden there is forced paternal leave and this has resulted in a narrowed income gap as women have more time to devote to career advancement. To me this suggests that parental roles play a large role in the career path of women.
If it is true that a smaller proportion of women than men have the aptitude to do math and science at the highest levels, there must be some actual evidence of it. Prior to the emergence of sex differences, say at the PhD candidate level in biology, what are the number of papers published per female PhD vs per male PhD ect.? Some data, or something similar, would be more convincing than looking at the sex ratio at higher positions in science.
BTW Kamel IS a man. now.

n/a said...

"Male" "PhD",

Ability to get a bachelor's degree does not equal ability to do high-level science. As you correctly point out, our current educational system seems to favor female traits. So what does it say that the vast majority of Nobel Prizes still go to men?

The "feminist graduate students" I was referring to are the coauthors of the paper. The "graduate student" part was a bit of a cheap shot, but I'm strongly of the belief that science should not be beholden to ideology.

rob,

As I said, sex differences should be taken into account where appropriate. However, the idea that the government should mandate that 50% of research subjects be female is outrageous. I assume that in most cases where authors don't report data separately by sex, they have already done basic checks and failed to find an interaction by sex. If males and females respond differently to a drug in important ways, for example, this should be apparent in a well-designed trial even if the proportion of research subjects who are female is 37% or 22% instead of 50%.

"However, these lifestyle choices are likely heavily influenced by societies traditional parental roles."

So what? Please watch the talk I linked to above. Feminism of the sort you and Kamel have been influenced by does not lead to either happier women or healthier societies.

"If it is true that a smaller proportion of women than men have the aptitude to do math and science at the highest levels, there must be some actual evidence of it."

Persistence of gender gap in top science and math performance. Or, try the entirety of human history.

n/a said...

"According to a recent article in
The Chronicle of Higher Education,
the Higher Education Research Institute
found that, as of 1989, 43%
of women in colleges and 20% in
universities had never published a
single journal article. The same was
true of only 23% of men in colleges
and 7% in universities
(Schneider
1998). Gender gaps in productivity
persist even when controlling for
educational origin, academic rank,
institutional type, and professional
age (Creamer 1998; Dinauer and
Ondeck 1999; Schneider 1998). In
addition, men continue to outpublish
women even in fields in
which women have been receiving
the majority of Ph.D.s for some time
(Creamer 1998; Schneider 1998).
Women are also less likely to be
highly prolific writers (writers who
account for a large proportion of the
literature in their field) (Creamer
1998). Men are three times more
likely to have published more than
10 journal articles than are their female
counterparts (Dinauer and Ondeck
1999).1 Creamer (1998) suggested
that prolific publishers are
disproportionately white males because
the career paths, work assignments,
research interests, and access
to resources conducive to frequent
publishing are more characteristic of
white men than of women and minorities."
http://www.saramitchell.org/mathewsandersen.pdf

Incidentally, the idea that high publication rates by white men reflect white male "privelege" is hilarious. The reverse is more nearly true: women and minorities are hired and retained with fewer or no publications.

n/a said...

Gender Differences in Patenting in the Academic Life Sciences
We analyze longitudinal data on academic careers and conduct interviews with faculty members to determine the scope and causes of the gender gap in patenting among life scientists. Our regressions on a random sample of 4,227 life scientists over a 30-year period show that women faculty members patent at about 40 percent of the rate of men. We find that the gender gap has been improving over time, although it remains large.
http://www.kauffman.org/uploadedFiles/gender_patenting_8306.pdf

rob said...

"I assume that in most cases where authors don't report data separately by sex, they have already done basic checks and failed to find an interaction by sex. If males and females respond differently to a drug in important ways, for example, this should be apparent in a well-designed trial even if the proportion of research subjects who are female is 37% or 22% instead of 50%."
I think that you trust pharmaceutical industry science more than I do. My cynicism would suggest that biases in sex testing are not done arbitrarily but to optimize positive results.

The numbers showing science as a male dominated occupation are not surprising due to the millions of factors Kamel mentioned. Again, lifestyle choices could account for these differences, meaning there maybe untapped scientific talent laying dormant in our population. The only way to utilize this resource is to ensure that lifestyle choices that affect career are truly free choices not negatively influenced by societial stereotypes and generalizations, of which you promote here. Perhaps I am biased against these gender generalizations since I know personally so many obviously exceptional female scientists, and perhaps, as you suggest, influenced by a feminist ideology, at its best an ideology of equality and tolerance. These influences may cloud an objective assesment, however they are clearly more constructive influences than the fear and hate I found when checking out your blog (especially some of the comments).
Classic stuff from i/a's blog: (which I won't link to)

(regarding a 1921 study on race differences in a fetus)

"The hand is relatively broader and slightly shorter and the thumb relatively longer in the white fetus than in the negro fetus"

Coincidentally, according to the second source I cite above (a review article from 2006):

"Non-human apes and brachiating monkeys have shorter thumbs and longer fingers, and also lower 2D:4D."

Hilarious stuff! Coincidentally indeed.
BTW: A recent article is not 1998 and I also enjoyed the link to data on seventh graders in response to my inquiry on data on PhDs. Cherry picking data in order to support a hypothesis you have decided is fact is somewhat deceitful.

i/a: Is there something that you propose should be done in light of the fact that women are less productive in science?

n/a said...

If pharmaceutical companies are putting out drugs that are dangerous to women without providing relevant warnings, they are liable. I don't recall seeing any class action law suits organized around that premise, so I'm going to have to assume it's not a major problem.

Reading a bit more about the senior author of the "sex bias" paper, I see that her conflicts of interest go beyond simply politics. Woodruff is involved in the:

Illinois Women's Health Registry, a growing database of 4,500 potential study subjects, that has encouraged Northwestern researchers to do more research in sex differences. About 1,000 of these women have participated in trials.

I think it's great that she's involved in something like that. I don't think it's so great that she's lobbying to have the government force researchers to make use of resources like hers.

"Is there something that you propose should be done in light of the fact that women are less productive in science?"

I don't see that anything needs to be done. Women who want to pursue science careers have more than adequate opportunity to do so. I don't believe in social engineering to try to force equality of outcome (nor is there any way such social engineering will lead to more or better science being done).

"I also enjoyed the link to data on seventh graders"

If you can't comprehend the relevance, I see no point in "debating" you.

"These influences may cloud an objective assesment, however they are clearly more constructive influences "

Having good intentions or being a good little boy who believes as he's told does not automatically make the ideas one promotes "constructive" or free one of moral culpability.

Kamel said...

MalePhD said: "62% of bachelor's graduates are women, how is that a small proportion. The times they are a changing."

That's a good point about the numbers. I only provided a one year (2006 - that was the latest year that data was available for all the stages I noted) snapshot. If you're interested in trends, I quickly put together a graph that you can see here. Again this is from NSF data, but only looking at Bio Sciences (Lumping together all science and engineering degrees, the numbers are considerably lower [eg. 50% B.Sc. and 33% postdoc]). There is a slow upward trend among Ph.Ds and postdocs, if real trends can be detected over a 6-7 year span.

Kamel said...

As for n/a, he said: Many women prefer "domestic roles". Many don't want to devote their lives as exclusively to work as men do.

Sure, except for the little fact that it looks like a higher proportion of female doctorates intend to go on to postdoc compared to men, yet in reality their numbers decrease.

"I'm not sure feminist graduate students are the best ones to be lecturing scientists on the topic"

Only scientists, a group that is predominantly male, are allowed to weigh in with their opinion? How convenient. What are your qualifications then?

I assume that in most cases where authors don't report data separately by sex, they have already done basic checks and failed to find an interaction by sex.

Why would you assume that? There was a clear example in the post. And if you follow the links, you'll find that only 20% of the studies looked at in the analysis mentioned why sex-specific data was not provided (the most common reason being
that differences by sex were found not to be significant). If they've "done the basic checks and failed to find an interaction by sex", it takes only one sentence in a paper to relay that information. If they're not saying it, why would you assume it?

If you have evidence women are paid less for the same work and level of experience, please present it.

I gave some evidence in the post, but yes it was broad categories and not narrow job descriptions. You can go to a site like payscale.com to see the sex differences by specific job (say for example, food scientist or research scientist, biotechnology or professor). They even make less at more female dominated professions like librarian or teacher or nurse.) Of course, you'll probably come back pointing out that it doesn't take into account years experience, and that's why their salaries are lower. Because you won't be happy unless it's the same job with the same experience at the same company in the same building.

BTW, did the studies you pointed to about how women are worse at science take into account the gender socialization that Rob and I mentioned? No? What a curious double standard!

Here's some of the work done for you: in countries with more gender-equal norms, the gender gap disappears.

Incidentally, did you even read the part of the paper you posted? The last sentence says it all: "prolific publishers are disproportionately white males because the career paths, work assignments, research interests, and access to resources conducive to frequent
publishing are more characteristic of white men." Access to resources is a 'characteristic of white men'? Hahahaha. Did you read the paper at all? I notice you didn't quote the part that says "Women are more likely than men to be excluded and isolated from the types of professional and social networks that define the life of a department" or "Subtle (and not-so-subtle) forms of discrimination include exclusion from institutional rewards like tenure and salary increases, sexual harassment, and a disregard for feminist or gender-related research." or any of the other parts that don't support your "they can't do science!" hypothesis. Cherry picking indeed! Should I even bother looking at the patenting paper to see how badly quote-mined it is?

Kamel said...

n/a said: If pharmaceutical companies are putting out drugs that are dangerous to women without providing relevant warnings, they are liable

This is a straw man. Nobody is saying the drugs are necessarily dangerous to women (but hey, don't you think that should actually be properly checked out before going to market?) They, or their dosing, may not be optimal. Again, why should women have less evidence-based care?

Kamel said...

Apologies for the last comment, yes there is some hinting at possible danger to improper gender considerations in trials: "This may be part of the reason why women are 1.5 times more likely to develop an adverse reaction to prescription drugs than men." So I guess my parenthetical stands: how do you propose to minimize adverse reactions without proper study? Do you have a better explanation for this difference? Saying it's not real because there's no lawsuit is ridiculous.

And that notwithstanding, the point about danger vs. optimization and evidence-based care still stands

Kamel said...

Not to keep picking at this thread, but this just occurred to me as mildly hilarious:

"I also enjoyed the link to data on seventh graders in response to my inquiry on data on PhDs."

If you can't comprehend the relevance, I see no point in "debating" you.


Earlier in the comments:

Ability to get a bachelor's degree does not equal ability to do high-level science.

It's true. Junior high is a MUCH better predictor.