There is a distinction that almost nobody in medicine can articulate precisely, and it explains a great deal about who succeeds.
A mentor talks to you. They meet you for coffee, review your CV, tell you what they would do, and offer genuine encouragement. It costs them an hour and nothing else.
A sponsor talks about you, in a room you are not in. They put your name forward for the fellowship. They say "she should give that talk" when the podium is being assigned. They call the chair at the other institution before your application arrives. They defend your candidacy when someone raises a concern.
Sponsorship costs the sponsor their own reputation. If you underperform, the person who put your name forward has spent capital badly, and everyone in that room remembers.
That is why sponsorship is scarce, and it is why the distribution of sponsorship, rather than the distribution of mentorship, determines who advances.
The evidence that medicine does not understand this is remarkable. In a 2026 survey of surgeons published in the American Journal of Surgery:
- 65 percent had a mentor.
- 28 percent had a sponsor.
- 17 percent had a coach.
- Only 48 percent could distinguish among the three.
Less than half of a surveyed surgical population could reliably tell the difference between someone who advises them and someone who spends capital on them.
Sponsorship follows power, not need
The same survey produced two findings that should reframe every institutional mentorship program in the country.
Community-based surgeons had mentor access at an odds ratio of 0.09 compared with academic peers. That is not a modest gap; it is near-exclusion.
And on sponsorship specifically: administrative leaders reported having sponsors at 44.8 percent, versus 21.3 percent for everyone else.
Read that carefully. The people who already hold institutional power are more than twice as likely to have someone spending capital on their advancement.
Sponsorship flows toward power rather than toward need, which is exactly backwards from how any deliberate system would allocate it, and precisely what you would expect from an entirely informal one. Sponsors sponsor the people they encounter, and the people they encounter are the people already in the rooms where sponsors are.
What the gap costs, measured
The consequences appear consistently across a series of large studies, and the magnitudes are not subtle.
Residency attrition. A JAMA Surgery meta-analysis covering 22 studies and 19,821 general surgery residents found overall attrition of 18 percent, with women leaving at 25 percent versus 15 percent for men (p = .008). Cited reasons included lack of role models, discrimination, and harassment.
Racial disparity in attrition. A study of 112,205 surgical residents from 2001 to 2018 found attrition of 10.6 percent among Black residents versus 6.2 percent among white residents, a relative risk of 1.66 overall and 2.59 for unintended attrition. In orthopaedics, the relative risk of unintended attrition reached 7.20.
Faculty promotion. An analysis of 31,045 academic surgical faculty from 2005 to 2020 found Hispanic women promoted at a hazard ratio of 0.68 and white women at 0.75, relative to white men. After ten years, 63 percent of underrepresented women remained versus 79 percent of non-underrepresented men.
Leadership composition. AAMC data for 2023 to 2024 puts women at 25 percent of department chairs and 27 percent of deans, with substantially lower representation in surgical specialties, at a time when women are the majority of medical students.
Each of these is well documented and frequently cited. What is less often connected is that they describe a cumulative pipeline effect in which the same mechanism operates at every stage: who is put forward, by whom, in rooms the candidate is not in.
The one study that measured the causal link
Most sponsorship discussion is theoretical. There is one dataset that measured it against outcomes, and it deserves far more attention.
Patton and colleagues, publishing in JAMA Internal Medicine in 2017, studied 995 NIH K award recipients, examining reported sponsorship against subsequent career success defined as obtaining an R01, publishing 35 or more papers, or achieving a leadership position.
Two findings:
- Sponsorship was more commonly reported by men.
- Sponsored recipients were significantly more successful, at p < 0.001.
That is, within a cohort already selected for early-career promise and funded by the same mechanism, sponsorship predicted who thrived, and it was unevenly distributed.
This is the single most important result in the field and it has generated far less institutional response than it warrants, largely because the obvious response, mandating sponsorship, is impossible.
Why institutions cannot fix this
Every structural feature of sponsorship resists institutional intervention, and understanding why is essential before proposing anything.
It cannot be mandated. An institution can assign a mentor. It cannot compel a senior physician to genuinely stake their reputation on someone. A mandated sponsor produces a name on a form and a perfunctory email, which is worse than nothing because it appears to solve the problem.
It cannot be observed. Sponsorship happens in appointment committees, in hallway conversations before conference programs are set, in phone calls between chairs. It is by definition invisible to everyone except the participants.
It cannot be requested. This is the cruelest constraint. A junior physician who asks to be sponsored has demonstrated precisely the lack of standing that makes sponsorship risky. The asking is itself disqualifying, which means the good is allocated entirely at the discretion of people who face no accountability for how they allocate it.
And it works best when unrecorded. The informal networks that distribute sponsorship efficiently among insiders function precisely because there is no accounting. Nobody tracks how many people any given senior figure has put forward, or who they have never put forward.
Because nothing is logged, sponsors face no reciprocity, no recognition, and no comparison. A senior surgeon who has sponsored fifteen junior colleagues over a decade receives no credit for it. One who has sponsored nobody faces no question. The behavior is therefore driven entirely by individual disposition and existing social proximity.
The mentorship program problem
Medicine's institutional response to all of this has been mentorship programs. Society matching schemes, departmental initiatives, structured pairings, dedicated leadership roles.
These are well intentioned and they address the wrong variable.
Look at the survey numbers again: 65 percent have a mentor. 28 percent have a sponsor. Mentorship is comparatively abundant. Sponsorship is scarce. Building more mentorship programs increases the supply of the thing that is not the constraint.
And there is a specific way this can make matters worse. An institution that runs a mentorship program can reasonably report that it has addressed professional development. The underlying sponsorship asymmetry continues untouched, now with a program obscuring it.
Meanwhile the institutional owners of this work have been contracting. Diversity-focused offices, which in many organizations were the closest thing to an accountable party for these disparities, have been substantially reduced across 2025 and 2026.
The problem has been precisely measured by a series of major studies and simultaneously orphaned. That combination is unusual and it is the reason to write about it now.
What would work: an accounting
If sponsorship cannot be mandated, observed, or requested, what remains?
Make it countable, and give the sponsor credit.
Consider what changes if a specific sponsorship act is recorded, with the protégé's consent, as a professional contribution: a nomination made, an introduction brokered, a letter written, a call placed, a name put forward.
Sponsors gain recognition for something currently invisible. The senior surgeon who has advanced fifteen careers has a record of it. This matters more than it sounds, because most sponsorship is done by people who are not recognized for it and who would do more if it counted as the professional contribution it plainly is.
Departments can see the distribution. Not to police individuals, but to answer a question no department can currently answer: who here has never been sponsored by anyone? That question is answerable from an accounting and unanswerable today, which is why the asymmetry persists until a cohort has already thinned.
Protégés get a legitimate route to ask. A structured, specific request ("would you make an introduction to X" or "would you nominate me for Y") is far more askable than the impossible general request to be sponsored. Specificity converts an imposition into a manageable favor.
And it creates reciprocity. A sponsor who has spent capital has a claim on the community's attention when they need something. Nothing currently closes that loop.
The supply is also identifiable. Senior and retired physicians have the most standing and the least to lose from spending it. A retired surgeon's nomination still carries weight in a department they no longer work in, and they face no competitive concern about advancing anyone. They are the most underused sponsorship capacity in medicine and nobody is organizing them.
What you can do now
If you are senior
Count your own sponsorship. In the last two years, whose name did you put forward, for what, in a room they were not in? Most senior physicians have never counted, and the exercise is uncomfortable in a useful way.
Then look at who is missing from that list. If everyone you have sponsored resembles you, that is not evidence of bias so much as evidence of proximity, and proximity is the mechanism the data describes.
Sponsor specifically, not generally. "I will mention you to the program committee" is a sponsorship act. "Let me know if I can help" is not, and the difference is the entire subject of this article.
Tell them you did it. Junior colleagues frequently do not know they have been sponsored, which means they cannot reciprocate, learn the mechanism, or eventually do it for someone else.
If you are junior
Learn the distinction and use it. You probably have mentors. Ask yourself who has actually put your name forward for something. If the answer is nobody, that is the gap, and it is a different gap from the one a mentorship program addresses.
Make specific, small, askable requests. "Would you introduce me to Dr. X" and "would you nominate me for that committee" are requests a senior person can grant in five minutes. "Will you sponsor me" is unanswerable.
Give your sponsor something to work with. A short paragraph they can paste into an email makes saying yes trivial. Most people who decline to sponsor are declining the work, not the person.
And notice the community-based penalty. The odds ratio of 0.09 for mentor access among community surgeons means that if you practise outside academia, you must build these relationships deliberately across institutional lines, because no structure will do it for you.
If you lead a department or a society
Stop counting mentorship and start counting sponsorship. Ask, once a year: for each junior faculty member, who has put their name forward for anything? The answer will identify the people about to leave.
Look at your own committee and podium rosters by who nominated whom. That is where sponsorship becomes visible, and it is data you already hold.
Recognize sponsors publicly. Making it a countable professional contribution is the single most available lever, because it addresses the reason the behavior is undersupplied.
Frequently asked questions
What is the difference between a mentor and a sponsor? A mentor gives advice and guidance directly to you. A sponsor advocates for you in rooms you are not in, spending their own reputation to advance your candidacy for opportunities. Survey research found 65 percent of surgeons had a mentor and 28 percent had a sponsor, with only 48 percent able to distinguish between mentorship, sponsorship, and coaching.
Does sponsorship affect career outcomes? The available evidence indicates it does substantially. A JAMA Internal Medicine study of 995 NIH K award recipients found sponsorship was more commonly reported by men and that sponsored recipients were significantly more successful on measures including R01 funding, publication volume, and leadership attainment, at p < 0.001.
What are attrition rates in surgical residency? A meta-analysis of 22 studies covering 19,821 general surgery residents found overall attrition of 18 percent, with women leaving at 25 percent versus 15 percent for men. A separate analysis of 112,205 surgical residents found Black residents leaving at 10.6 percent versus 6.2 percent for white residents, with 2.59 times the relative risk of unintended attrition, rising to 7.20 in orthopaedics.
Are promotion rates equal in academic surgery? No. Analysis of 31,045 academic surgical faculty from 2005 to 2020 found Hispanic women promoted at a hazard ratio of 0.68 and white women at 0.75 relative to white men, with ten-year retention of 63 percent among underrepresented women versus 79 percent among non-underrepresented men.
Why don't mentorship programs close these gaps? Because they increase the supply of the abundant good rather than the scarce one. Mentorship is comparatively available at 65 percent; sponsorship is scarce at 28 percent and concentrated among those already holding institutional power, with administrative leaders reporting sponsors at 44.8 percent versus 21.3 percent for others.
Can sponsorship be institutionalized? Not by mandate, since a compelled sponsor produces a form entry rather than genuine reputational advocacy. What can be institutionalized is accounting: recording specific sponsorship acts with consent, giving sponsors credit for a contribution that is currently invisible, and allowing departments to identify who has never been sponsored by anyone.
The bottom line
Medicine has measured this problem exceptionally well. We know surgical residency attrition runs 25 percent for women against 15 percent for men. We know Black residents face more than two and a half times the risk of unintended attrition. We know underrepresented women faculty are promoted more slowly and retained less. And we know, from the one study that looked, that sponsorship predicts success and is unevenly distributed.
The response has been mentorship programs, which supply the abundant good, while the scarce one continues to flow toward people who already have institutional power at more than twice the rate.
Sponsorship cannot be mandated, cannot be observed, and cannot be asked for without the asking undermining the request. It happens in rooms with no minutes, performed by people who receive no credit, on behalf of people who frequently never learn it happened.
The only thing nobody has tried is counting it. Not to police anyone, but because a good that is invisible cannot be distributed deliberately, and because the senior physicians who do this work generously right now have no idea they are the exception.
Part of a series on the missing professional infrastructure of healthcare. Previously: Your Colleagues Are Strangers
Evidence note: sponsorship prevalence and access figures come from a 2026 American Journal of Surgery survey with a sample of 93, which is small and should be read as indicative. Attrition figures come from Khoushhal et al. in JAMA Surgery (2017) and a JAMA Surgery analysis of 112,205 surgical residents (2023). Faculty promotion and retention figures come from JAMA Surgery (2024) covering 31,045 faculty. Sponsorship and career outcomes come from Patton et al. in JAMA Internal Medicine (2017). Leadership representation figures come from AAMC State of Women in Academic Medicine 2023-2024. Observational studies of attrition and promotion cannot establish that sponsorship differences cause the disparities described, and multiple mechanisms are documented in this literature.
