The Gender Data Gap in Clinical Trials

In 1977, the FDA drew a line and told drugmakers to keep women of "reproductive potential" out of Phase I and Phase II trials. The policy was a direct response to thalidomide, the sedative that caused severe birth defects across Europe and Canada, and the intent was to protect fetuses from an untested drug. The effect, though, was to bar an entire sex from the exact trials where dose safety and metabolic behavior get worked out in the first place, the foundational layer everything else in drug development sits on top of.

That exclusion didn't stay contained to whatever the FDA originally had in mind. As the AIDS epidemic grew through the early 1980s, the same logic kept women out of HIV drug trials, at a moment when access to experimental treatment was often the difference between life and death. Patient activists eventually forced the issue into public view, and the policy stayed on the books until 1993, when federal law finally required the inclusion of women and minorities in federally funded research. Sixteen years is a long time to build a body of drug knowledge with half the population missing from the blueprint, and the 1993 law was supposed to close that gap. It didn't, and the rest of this piece is about why.

Where enrollment actually stands three decades after the legal mandate to include women

A 2022 study in Contemporary Clinical Trials looked at 302,664 participants across 1,433 U.S. trials run between 2016 and 2019. Women made up just 41% of enrollment on average, and that number held steady across both cardiovascular and cancer trials.

Psychiatry shows the gap even more starkly. Women account for 60% of people living with psychiatric disorders, yet they make up only 42% of participants in psychiatric trials. Nephrology tells a similar story: a meta-analysis of trials from 2000 to 2021 found men made up 62% of the 215,850 people enrolled, women only 38% (133,082 people). Break it down by condition and men were the majority in acute kidney injury trials (65%) and dialysis trials (55%), with similar male-skewed patterns across other nephrology subfields.

Then there's obstetrics, which might be the clearest case of the mismatch. Obstetrical complications affect more than a third of women worldwide, yet obstetrical trials make up about 2% of all U.S. clinical trials. Sit with that ratio for a second: a condition touching one in three women gets a sliver of the research pipeline.

The easy explanation is that women decline to sign up, but the evidence points elsewhere. Structural features of trial design appear to shape who ends up enrolled, reflecting choices baked into the design of trials rather than the preferences of the people being recruited for them. The mandate changed the law in 1993, but it didn't change the trial design playbook.

Why the gap is worse in early-phase trials, where drug safety is actually established

Here's where it gets uncomfortable: only 30.6% of Phase I trial participants are women, and 34.1% of early-phase trials enroll exclusively men. Zero women — not underrepresented, just absent.

Phase I is where researchers figure out dose tolerability, how a drug moves through the body, and what side effects show up first. It's the trial phase that answers the question "how much of this can a person safely take?" If women aren't in the room for that question, the answer defaults to male physiology, and everything downstream, every later trial phase, every approved label, every pharmacist's dosing chart, inherits that default.

The problem goes back even further than human trials. Roughly 80% of non-clinical animal studies use only male animals, so sex differences in drug response, the kind that might show up in a mouse before they show up in a person, never even get the chance to appear. Why? Researchers have historically worried that fluctuating hormones in female animals would "confound" the results, a rationale that quietly treats the male body as the neutral, uncomplicated baseline and the female body as statistical noise to be avoided rather than data to be collected.

The NIH actually tried to fix this. In 2016, it rolled out a policy called Sex as a Biological Variable, or SABV, which required scientists to account for sex in both animal and human studies; anyone who wanted to run a single-sex study had to justify why. It was a real requirement with real teeth, for a while, but then, in early 2025, NIH quietly relabeled its SABV policy pages as "Historic Documents," filed under history rather than formally repealed or debated. So the preclinical and early-phase pipeline, the two stages furthest from public visibility and furthest from FDA-approval headlines, is exactly the mechanism through which male-default assumptions get baked into every stage that follows. Nobody's cutting a ribbon on the mouse study, but that's where the die gets cast.

What different biology actually does to drug behavior in women's bodies

This is a story about metabolism, and metabolism doesn't read op-eds or care about mandates. Enzyme activity in the liver's CYP family, the density of drug transporters, circulating sex hormone levels, these differ between men and women in ways that change how a drug gets absorbed, broken down, and cleared.

One pharmacokinetic study looked at 86 drugs and found that 76 of them produced higher pharmacokinetic values in women: higher blood concentrations, longer elimination times, the drug simply sticking around longer or hitting harder. Of 59 drugs with clinically identifiable adverse reactions tied to sex, the pharmacokinetic sex bias predicted the direction of the adverse reaction bias in 88% of cases. That's not a loose correlation; that's close to a rule.

A separate look at drug exposure found 25 drugs where the sex difference in exposure was large enough to matter clinically, defined as at least a 50% change, and that difference tracked with how well the drug actually worked for one sex versus the other. Put plainly: a standard dose worked out mostly from male trial data is a male dose, calibrated to a male body, applied to a body it wasn't designed for, and calling it "standard" doesn't make it universal.

Zolpidem as the case study: what happens when sex-specific pharmacology is ignored until after approval

Ambien is the drug that makes this whole argument concrete instead of theoretical. It got approved at a single dose for everyone, and years passed before adverse event data kept piling up showing women experienced significantly more next-morning driving impairment than men taking the identical dose, meaning women were getting behind the wheel still functionally sedated while the label said nothing about it.

In 2013, the FDA finally issued a Drug Safety Communication requiring a sex-differentiated dose: 5 mg for women, 5 or 10 mg for men. That's a 50% cut for women, which tells you something about how far off the original dose was.

Here's the part that should bother you more than the correction itself: Ambien remains the only FDA-approved drug carrying sex-differentiated recommended doses, one drug out of the entire pharmacopeia. The pharmacokinetic data in the last section suggests dozens of other drugs carry a similar mismatch; Ambien is simply the one where the mismatch got loud enough, through enough car accidents and adverse event reports, to force a label change. The biology was there from day one, sitting quietly in blood concentration curves, and it just took millions of women taking the wrong dose before anyone went looking for it. The fix arrived after the fact, not before it.

How the adverse drug reaction data shows the aggregate cost across all approved medicines

Zoom out from any single drug and the pattern holds at scale. Women report serious adverse events from approved medicines 52% more often than men, and serious or fatal events specifically, 36% more often. The FDA's Adverse Event database logged 12.9 million reports involving women through 2022, versus 8.5 million for men.

So which is it: are women just more likely to report a problem, or more likely to actually have one? The pharmacokinetic evidence from the last two sections points pretty hard toward the second explanation. If this were purely a reporting behavior gap, driven by women engaging more with the healthcare system generally, you'd expect that gap to show up evenly across mild, moderate, and serious events alike. It doesn't; it concentrates specifically in the serious and fatal categories, exactly where you'd expect a real physiological mismatch to surface rather than a reporting quirk. That's the aggregate scorecard. The next section breaks it down by disease, and it's not any prettier close up.

Disease by disease: where the enrollment gap produces the clearest downstream harm

Cardiovascular disease kills more women than any other cause of death, and yet it spent decades being framed in research circles as a disease of men in their fifties clutching their chest. Angioplasty, coronary stents, the standard toolkit, got developed and refined on trial populations that skewed male, and women have been found to experience adverse outcomes after certain cardiovascular procedures at higher rates than men. Add to that the guidelines built around a textbook chest-clutching presentation, while women's cardiac symptoms may differ from the textbook presentation and get systematically underweighted or missed entirely by clinicians trained on the male script.

Mental health has its own blind spot. Fewer than 1% of the 768 depression treatment trials analyzed in a study published on pmc.ncbi.nlm.nih.gov even planned to analyze results by gender. Women make up 60% of people with psychiatric diagnoses and 42% of psychiatric trial participants, so the evidence base treating these conditions is tilted away from the population that carries most of the burden.

Oncology, despite being one of the most heavily funded areas of medical research, has the same hole in it. Of 89,221 oncology trials, only 472, about half a percent, had curated sex comparisons after treatment. Among the 288 trials that did make the comparison, 42% found significantly better outcomes for women, 16% for men. That's a real biological signal, sitting right there, and almost nobody built a trial designed to catch it.

Sexual and reproductive health drugs show the research priority gap in its bluntest form: as of 2024, men had 26 FDA-approved sexual health drugs on the market, while women had 2. Female sexual and reproductive physiology isn't simpler than male physiology; it's just been researched less.

Even COVID-19, a global emergency with an enormous, fast-moving research investment, didn't escape the pattern. A systematic review of 75 clinical trials found only 24% presented outcome data broken out by sex, and only 13% discussed what the differences might mean separately for men and women. If a once-in-a-century pandemic response couldn't consistently disaggregate by sex, it's hard to argue the infrastructure for doing so is anywhere close to routine.

Why the gender of trial leadership correlates with how many women are enrolled

Here's a pattern worth sitting with: research has found that the gender of trial leadership correlates with the share of women enrolled, with female-led trials tending to enroll more women than male-led trials, a pattern that persists even after adjusting for other factors.

A separate analysis tracking almost 160,000 trials submitted to ClinicalTrials.gov from 2010 to 2023 found female PI representation climbing from around a third in 2010 to 41% by 2023, real progress on paper. But break it down by field and cardiovascular trials, the exact category with the most documented harm to women, have lagged behind other specialties in female PI representation.

Correlation isn't proof of causation here, and it's worth saying so plainly: women PIs may simply gravitate toward studying conditions that disproportionately affect women, which would produce this same pattern without anyone's leadership gender directly driving enrollment decisions. But the correlation survives controlling for trial type, which suggests something more than self-selection is going on. Who sits at the table when a trial gets designed seems to shape who ends up walking through the door to join it. That makes the pipeline of women into research leadership a question bigger than representation for its own sake.

The regulatory progress made between 1993 and 2024, and what has been reversed since

Trace the actual arc and it looks like real, if slow, momentum: the 1993 NIH Revitalization Act, then the creation of the FDA Office of Women's Health, then the 2016 SABV policy, then the 2022 Food and Drug Omnibus Reform Act, which required trial sponsors to submit Diversity Action Plans alongside late-stage trial protocols. FDA draft guidance followed, intended to spell out how those diversity plans should actually work in practice.

Then, shortly after a new administration took office in 2025, the FDA quietly pulled that draft guidance. Around the same time, as mentioned earlier, the NIH filed its SABV policy under "Historic Documents." Neither move came with much fanfare, yet both amount to the same thing: a rule got replaced by a suggestion.

That distinction matters more than it sounds like it should. A mandatory reporting requirement means sponsors have to show their work and face consequences if the work isn't there. Voluntary guidance means sponsors can self-report their diversity efforts with nobody checking the math. The pattern across three decades is consistent: every step forward took years of advocacy, usually with a crisis attached (thalidomide, AIDS activism, adverse event data piling up), while every step back seems to take a signature and a quiet policy-page edit.

The economic argument for closing the gap, and why it has not moved faster

If the moral case hasn't been enough to move the needle in three decades, maybe the money will. A 2024 report from the McKinsey Health Institute and the World Economic Forum estimated that closing the gender health gap could add up to $1 trillion annually to the global economy by 2040, cut the time women spend in poor health by nearly two-thirds, and hand back roughly seven additional healthy days per woman, per year.

Right now, women spend 25% more of their lives in poor health than men, adding up to an estimated 75 million disability-adjusted life years lost every year, and less than 2% of healthcare research investment goes toward conditions specific to women, beyond cancer.

Here's the part worth pausing on: if this gap were purely a market failure, some drug company somewhere would have already found the trillion dollars sitting on the table and gone after it. The fact that it's still sitting there after decades suggests the problem is also a knowledge failure, running alongside the economic one. When male-default data gets treated as universal data, there's no visible gap for anyone to notice, let alone fill, and nobody chases a market they don't believe exists.

What the trillion-dollar number does usefully is translate the whole argument into a language that reaches decision-makers who don't respond to fairness arguments on their own. But the underlying fix has to be bigger than a bigger number of women showing up to enroll. Bad data at the animal-study stage produces biased Phase I trials, which produce underpowered subgroup analyses, which produce treatment guidelines that don't reflect how female bodies actually process the drug, which produces avoidable adverse events, the kind sitting in that 12.9-million-report pile. Fixing enrollment numbers without fixing that chain is necessary, but it just isn't sufficient on its own.