For years, the standard explanation for why so few women lead health organisations went like this: leadership reflects the workforce of twenty or thirty years ago. Wait for the women now in training to rise, and the numbers will even out on their own.
The waiting has been going on for a long time.
In 2019, women became the majority of students in US medical schools for the first time, at 50.5 per cent, according to the Association of American Medical Colleges. But women had made up at least 45 per cent of medical school classes for about two decades before that. By then, the AAMC reported, women were still only 25 per cent of full professors, 18 per cent of department chairs and 16 per cent of medical school deans.
A pipeline that has been nearly half women for twenty years and still produces leadership that is four-fifths men is not a pipeline problem. It is a leak, and the leak has specific locations.
The global picture: delivered by women, led by men
The pattern is not specific to American academic medicine. The WHO’s 2019 report on gender in the health workforce was titled, bluntly, “Delivered by women, led by men.” Women make up almost 70 per cent of the global health and social care workforce and close to 90 per cent of nursing and midwifery, yet hold around 25 per cent of leadership roles in health, according to Women in Global Health and WHO analyses. Women in the sector earn, on average, 24 per cent less than men, according to a 2022 WHO and International Labour Organization report.
The people who do most of the work in health systems are the people least likely to decide how those systems are run.
Where the leak actually happens
The popular image of the problem is a glass ceiling: a barrier at the very top, facing women who reach senior levels. The data increasingly point somewhere else, much earlier.
McKinsey and LeanIn.Org’s Women in the Workplace 2024 report, covering sectors broadly, found that for every 100 men promoted to their first manager role, only 81 women were promoted. The authors call this the “broken rung.” Because fewer women make the first step, there are fewer women at every level above it, and no amount of attention to executive recruitment can fully fix a shortage that was created at the bottom.
The rung is more broken for some women than others. The same report found that for every 100 men promoted to manager in 2024, 54 Black women and 65 Latinas were promoted. Any analysis of health leadership that looks only at gender, and not at who among women is advancing, will overstate progress.
In health, the equivalent moments are recognisable. Who is nominated to chair the first committee. Who gets protected time for research while others carry the clinical load. Who is sponsored for the first department role. These decisions are small, informal and rarely tracked, which is exactly why they compound.
How small gaps become large ones
It is worth doing the arithmetic, because it shows why patience never worked.
Imagine an organisation that starts with equal numbers of men and women at entry level. Suppose that at each promotion step, women advance at 81 per cent of the rate men do, the ratio McKinsey found at the first rung, and suppose for simplicity the same ratio applied all the way up. This is an illustration, not a model of any real institution, but the effect is striking.
After one step, women are about 45 per cent of the cohort. After three steps, about 35 per cent. After five, around 26 per cent. No single decision looks dramatic. A promotion committee that picks a woman four times in nine would hardly be accused of bias. Yet five such steps, from junior doctor or nurse to the top of a hospital or university department, are enough to turn an equal intake into a leadership group roughly three-quarters male. That is almost exactly what the AAMC and WHO figures describe.
The illustration also shows why the pipeline argument fails on its own terms. A bigger intake of women raises the starting point, but if the leak at each step stays the same, the top still ends up heavily skewed. Doubling the water in a leaking pipe gets more water out of the end. It does not stop the leak.
The arithmetic cuts the other way as well, which is the hopeful part. Because the losses compound, small repairs compound too. Lift the ratio at each step from 81 to 95 women per 100 men and the same five-step ladder ends with women at around 44 per cent of the top group rather than 26. No quota is required, only a slightly fairer decision at each of the ordinary moments where careers are shaped. That is a far more achievable goal than transforming a chief executive search, and it is where the returns are largest.
The uncomfortable trade-offs
Some explanations for the gap point to women’s choices, especially around caregiving. Those choices are real, but they are not made in a vacuum. Leadership roles in health are often designed around unlimited availability: evening meetings, relocations, long unpaid committee apprenticeships. Anyone with caregiving responsibilities, who are still disproportionately women, is quietly filtered out.
There is also an incentive problem inside organisations. Leadership teams recruit people who resemble them and who are already known to them. That is not always conscious bias. It is the path of least resistance. Changing it requires deliberate effort that takes time and carries some political risk for the people who push for it.
And the incentives run the wrong way at exactly the wrong moment. The informal work that often falls to women early in their careers, such as teaching, mentoring juniors, sitting on wellbeing or equality committees, is valued in appraisals as good citizenship but rarely counts towards promotion. The work that does count, such as grants, publications, service-line growth, is easier to do for someone whose time is protected. A system can praise women for carrying its institutional load and then promote the colleagues who were freed from it.
The strongest counterargument
The best version of the other side goes like this. Leadership roles demand long hours and total availability because the work genuinely requires them. If more women than men choose, freely, to prioritise other things at certain points in their lives, a lower share of women at the top may reflect preferences rather than barriers. Forcing equal outcomes could mean appointing people who want the job less.
Part of this holds. Preferences differ between individuals, and no one should be pushed into a role they do not want. Some senior jobs will always be demanding.
But the argument has two weaknesses. First, the broken rung appears at the first managerial step, early in careers and often before caregiving pressures peak, which is hard to explain by preference alone. Second, “the job requires unlimited availability” is frequently a description of how the job has been designed, not a law of nature. Many of the evening meetings and unpaid apprenticeships exist because the people who designed them never had to work around them. When roles are redesigned, the preference argument usually shrinks, which suggests it was partly measuring the design all along.
Why it matters beyond fairness
Fairness would be reason enough. But there is also a practical case. A health system that promotes from a narrower pool than it trains is wasting talent it has already paid to develop. Leadership that reflects the workforce is better placed to understand the working conditions that drive burnout and turnover, the realities of frontline care and the needs of the patients those systems serve.
There is a second-order effect, too. Nursing, midwifery and community health work are overwhelmingly female, and their perspective is thinly represented in the rooms where budgets and strategies are set. When those rooms are dominated by one profession and one gender, decisions about staffing, scheduling and workforce investment are made at one remove from the people who carry them out. The leadership gap and the retention crisis are not separate problems.
What actually moves the numbers
The most effective interventions focus on the specific points where women are lost, and on accountability rather than aspiration.
Start by tracking the first promotion. Measure who is promoted to first leadership roles, by gender and ethnicity, every year, and publish it internally. What gets counted at the first rung gets managed there. Pair that with sponsorship, not just mentoring. Mentors give advice; sponsors put names forward for roles and spend their own credibility doing so. Many women in health are over-mentored and under-sponsored.
Make the criteria for leadership roles transparent, because informal appointments favour existing networks. Redesign roles around outcomes rather than availability: meeting times, flexible arrangements and job-sharing at senior levels widen the field without lowering the bar. Fix pay gaps where they are found, since audits only matter if they lead to adjustments. And hold leaders accountable. Diversity targets without consequences are wishes.
What this means for specific actors
For hospital chief executives and medical school deans, the practical question is whether they can say, today, what share of first-line leadership appointments went to women last year and how that compares with the eligible pool. If not, the first job is to find out.
For boards and governing bodies, succession planning is where the leak becomes visible. A board that only sees a shortlist at chief executive level is seeing the end of a fifteen-year process it had no hand in. Asking to see the pipeline two or three levels down changes what executives pay attention to.
For ministries of health, the lever is the public system itself, often the largest health employer in the country. Publishing leadership data for public hospitals and national agencies, and including it in performance frameworks, sets a standard the private sector tends to follow. For funders and research councils, grant rules on protected time, parental leave extensions and panel composition shape who gets the track record that leads to leadership. For professional bodies and royal colleges, their own councils and committees are a visible test of whether they mean it.
What to watch in the next two years
The signals worth following are specific. Watch whether future AAMC data show the share of women among department chairs and deans closing on the share among full professors, which would suggest the upper rungs are being repaired. Watch the next Women in the Workplace report for any movement in the broken rung, especially for women of colour. Watch whether WHO and Women in Global Health continue to publish leadership data that allow country comparisons, since visibility has been the main driver of change so far. And watch individual health systems: the ones that begin publishing first-promotion data are the ones most likely to be different in a decade.
The takeaway
The pipeline argument was always comforting because it asked nothing of anyone except patience. Twenty years of women filling medical school classes, and nursing staff that is overwhelmingly female, have tested it. The pipeline is full. The leaks are in how health organisations choose, sponsor and promote their leaders, and those are choices that can be changed now.
Where have you seen talented women drop out of the leadership track in your organisation, and what could have kept them? Tell us in the comments.
