It is nine at night and a general surgeon is sitting at her kitchen table with a retractor she has modified three times with a Dremel tool and a 3D-printed prototype clip, trying to solve a problem that has bothered her for six years: the standard instrument for a specific step in a common procedure is wrong, and she has known exactly how to fix it since her third year of practice.
She filed an invention disclosure with her hospital's technology transfer office two years ago. It sits in a queue behind device ideas from cardiology and orthopedics, departments with more patent history and, not coincidentally, more relationships with the office's small staff. Nobody has called her back with a real answer about next steps. She does not know an engineer who works in her instrument's specific modality. She does not know anyone who has been through the FDA's 510(k) clearance process and can tell her what actually happens after you submit. She does not know what her hospital's employment contract actually entitles her to if the device ever earns a dollar.
She mentioned the idea, once, to a device sales rep who visits her operating room monthly. He nodded with real interest and asked good questions. Eight months later, a nearly identical clip appeared in a competitor's catalog, filed by an engineer she had never met.
She had the idea a working device needed. What she did not have, and could not find anywhere in her own hospital or her own professional network, was the person who could turn it into something she owned.
The number that reframes the whole problem
Start with what the patent record actually shows, because it overturns the obvious assumption.
Between 1993 and 2018, the US Patent and Trademark Office issued 275,260 surgery-related patents. Surgeons held 9,008 of them, just 3.3%, and that thin slice was itself concentrated in 2,164 individuals, roughly 4% of American College of Surgeons Fellows. Of those 2,164 patent-holding surgeons, 97% were male.
Over the same 25-year window, surgical patenting grew 462%. Surgeon participation in that growth barely moved.
Read those two facts together and the obvious explanation, that surgeons simply do not have good ideas at the rate other inventors do, falls apart immediately. Surgeons stand at the exact point in the healthcare system where instrument failure is most directly, physically obvious: they are the ones holding the wrong tool. The people closest to the problem are the ones least represented among the people who solved it on paper. The invention rate is not the constraint. The path from observation to filed, cleared, licensed device is.
What that path actually costs, and where it breaks
Device development is not a fast or cheap process under the best conditions. It is commonly quoted at three to ten years and $31 million to $94 million from initial concept through market. A practicing surgeon working nights and weekends on a kitchen-table prototype is never going to personally fund that. She needs partners, and the partners are where the system breaks.
The most common regulatory route for a device like the one in the opening scene is the FDA's 510(k) pathway, which clears a device by showing it is substantially equivalent to something already on the market. In 2020, across more than 3,000 applications, the average refuse-to-accept rate was 32.43%, meaning roughly a third of submissions were rejected before even reaching substantive review, typically for administrative or completeness failures rather than scientific ones. Within certain CDRH offices, the average refuse-to-accept rate climbed to 46%, with rejection rates over 60% inside some offices.
Each refuse-to-accept letter is not a minor setback. It is months of a clinician-inventor's limited spare time and tens of thousands of dollars of avoidable spend, almost always because the submission was assembled by someone who had never actually been through the process before and did not know the specific, unwritten expectations of that particular office. A surgeon who has already cleared a device through that exact CDRH office is worth, in practical terms, more than any amount of general regulatory consulting, because they know precisely what that office wants to see and precisely what will trigger a rejection.
The institution that is supposed to help does not have the capacity. University technology transfer offices, according to a 2023 analysis in the Journal of Clinical and Translational Science, "are resource-limited in how far and to what degree they can assist innovators," and most universities have limited regulatory expertise in-house at all. The office holding the surgeon's disclosure is, by its own field's honest self-assessment, not equipped to move it forward on its own.
The pipeline that works, and how small it is
There is proof this can be done well. It is just concentrated in a handful of zip codes.
Stanford's Byers Center for Biodesign has trained 243 Innovation Fellows since 2001 and put more than 4,000 students through its broader programs since 2002. That is a genuinely serious, decades-long pipeline, roughly the size of one residency program's total alumni network, and it has produced real devices and real careers.
It is also, by design, small, elite, and geographically concentrated. A surgeon in a community hospital outside the small set of institutions running a program like Biodesign has no equivalent structured path at all. She has her own initiative, a Dremel tool, and a technology transfer office that has told her, honestly, that it does not have the regulatory expertise to help her very far.
Why nobody owns this
Look at every party positioned to close the gap, and each one has a structural reason it has not.
Tech-transfer offices are measured, and funded internally, based on revenue from a small number of blockbuster licenses, not on how many disclosures they successfully move forward. A surgeon's promising-but-unproven idea does not move the metric the office is actually managed against.
Device companies profit when a surgeon's idea arrives as informal input rather than as protected, licensable intellectual property. The sales rep in the opening scene had every incentive to listen closely and no incentive to help the surgeon file first.
Medical societies run innovation sessions as conference programming, panels, talks, the occasional pitch competition, not as an ongoing matching service that connects a specific surgeon to the specific engineer and regulatory veteran she actually needs.
Biodesign-style fellowship programs are excellent and deliberately small. They were built to train a cohort deeply, not to serve the much larger population of practicing surgeons who already have an idea and simply need a path, not a fellowship year away from clinical practice.
Nobody is paid to connect the 96% of surgeons outside the patent-holding pool to the 4% inside it. That is the entire gap stated plainly: the people who already know how to do this successfully are identifiable, they are colleagues within the same profession, and there is no mechanism routing a new inventor to them.
The structural failure: three people, and no channel to any of them
The actual bottleneck, stated precisely, is a short list. A clinician-inventor needs an engineer who works in the relevant modality and can build a real prototype. She needs a peer who has personally been through a refuse-to-accept letter at the specific CDRH office her device will land in, and knows what that office actually wants. And she needs a licensing officer, inside her own institution or at a comparable one, who will say plainly what terms are actually available, rather than what a generic policy document implies.
Those three people all exist inside medicine's own professional network, right now, distributed across thousands of hospitals. Patent databases record who eventually succeeded. They record nothing about the path: who almost gave up at month eight, who found the right engineer through a lucky introduction, who got a refuse-to-accept letter and knew exactly how to fix it because they had already seen the same letter once before. Tech-transfer offices know only their own institution's file. Society meetings showcase finished successes at the podium and structurally hide the failures that would carry the actually transferable lesson.
This is a done-it-before graph that does not exist. Every other version of that same failure documented across this series, the surgeon with no one to call after a complication, the investigator with no way to know a study was already tried and abandoned, shows up here in a commercial, higher-dollar form: the knowledge that would unblock a stalled invention is sitting in a specific colleague's memory, and there is no channel connecting that memory to the person who needs it.
What would actually work
A structured intake routed to verified, attested experience, not a general innovation newsletter. Problem, modality, prior art status, and institutional employment terms, matched to members who have actually filed, cleared, or licensed something comparable, not simply to whoever volunteers.
Ownership and confidentiality terms settled before any disclosure happens. A surgeon should know, going in, exactly what she is agreeing to share and what protection she has, rather than negotiating that after an idea has already left her hands informally, the way it did with the sales rep in the opening scene.
A retired and 55+ surgeon population treated as core supply, not an afterthought. Surgeons who have already cleared devices and are past their own institutional non-compete and employment constraints are exactly the underused resource this problem needs, and there are more of them, at every career stage, than there are active Biodesign fellows in any given year.
A path dataset that tracks the whole route, not just the outcome. Which specific pathway worked, by specialty, by institution type, by CDRH office, built from real cases rather than only the patents that eventually issued, would let a new inventor see roughly what to expect before she starts.
Community-employed clinicians treated as an explicit target, not an edge case. The current system is heavily concentrated at a small number of elite, well-resourced institutions; a serious fix has to work for the surgeon whose hospital has no research office at all.
Compliance built into the structure, not bolted on. Conflict-of-interest rules when a surgeon selects a device for her own patients, Sunshine Act reporting on royalties, and informed-consent obligations when a surgeon uses a device she herself invented all need to be addressed as part of the design, not discovered after the fact.
What you can do now
If you are a clinician with an unbuilt device idea
File the disclosure with your institution's tech-transfer office anyway, even knowing its resource limits. It establishes a documented priority date and a paper trail, and costs you almost nothing beyond the time to write it up clearly.
Look specifically for a peer who has cleared a device in your exact specialty, not just any physician-inventor. A general innovation mentor is far less valuable than someone who has personally been through a refuse-to-accept letter in your modality.
Read your employment contract's intellectual property clause before you talk to anyone outside your institution about the idea. Knowing what you actually own, and what your hospital is entitled to, changes every conversation that follows, including the one you have with a device sales rep.
If you are a retired or semi-retired surgeon who has cleared a device
Make yourself findable as someone willing to advise on a specific, narrow question: what happened when you cleared your device. That single conversation, the one the surgeon in the opening scene never had access to, is worth more to a first-time inventor than almost any general mentorship.
Be explicit about which CDRH office and which pathway your experience covers. Specificity is what makes an introduction actually useful rather than a general encouragement to keep going.
If you run a technology transfer office
Say plainly, to your own disclosing clinicians, what you can and cannot help with. The 2023 finding that most offices have limited in-house regulatory expertise is not a secret failure; naming it honestly, early, saves an inventor months of assuming help is coming that is not.
Build a referral relationship to peer clinician-inventors outside your own institution, deliberately. You do not need in-house regulatory expertise if you have a reliable way to route a disclosure to someone who has already solved the exact problem it presents.
Frequently asked questions
What percentage of surgical patents are held by surgeons? Just 3.3%. Between 1993 and 2018, the USPTO issued 275,260 surgery-related patents, of which surgeons held 9,008, concentrated in 2,164 individuals, about 4% of American College of Surgeons Fellows, 97% of them male (Annals of Surgery, 2022).
Can a hospital-employed physician own their own invention? It depends entirely on the specific employment contract and institutional intellectual property policy, which vary widely between academic medical centers and community systems. Physicians are strongly advised to review their contract's IP clause before disclosing an idea to anyone outside the institution.
What is a 510(k) refuse-to-accept letter? It is FDA's rejection of a device clearance submission before substantive review begins, typically for incompleteness or administrative deficiencies. The average refuse-to-accept rate across more than 3,000 applications in 2020 was 32.43%, rising to an average of 46% and rejection rates over 60% within some individual CDRH offices (Journal of Clinical and Translational Science, 2023).
How long does it take to bring a medical device to market? Commonly quoted at three to ten years from concept to market, with total development cost estimated at $31 million to $94 million (Journal of Clinical and Translational Science, 2023). Most of that time and cost falls on parties other than the original clinician-inventor once real development begins.
How do surgeons get royalties on devices they help design? Royalty terms are set by the licensing agreement between the inventor (or their institution, if the invention arose from institution-supported work) and the manufacturer, and vary widely; there is no standard rate. Physicians receiving device royalties are also subject to Sunshine Act reporting and must navigate anti-kickback and Stark Law considerations if they use their own royalty-bearing device on patients.
Where can a surgeon get help patenting a device? The formal starting point is a hospital or university technology transfer office, but a 2023 analysis found most such offices are resource-limited and most universities have limited in-house regulatory expertise, meaning many disclosures stall for lack of engineering and regulatory partners rather than lack of institutional process.
The bottom line
The surgeon at her kitchen table is not an outlier. She is close to the median experience of the tiny fraction of surgeons who ever attempt to formalize an idea at all, which is itself a tiny fraction of the surgeons who have one. Ninety-seven percent of the surgeons who successfully hold a patent are men, concentrated at a handful of institutions, out of a profession that is neither that homogeneous nor that geographically narrow. The idea supply is broad. The path is not.
The patent record proves the constraint is not creativity: surgical patenting grew 462% over 25 years while surgeon participation in that growth barely moved, which means the growth came from everyone standing between the surgeon and the patent office, engineers, device companies, non-clinician inventors, rather than from the surgeons themselves. Meanwhile the one national pipeline built to close this gap, Stanford's Biodesign program, has trained 243 fellows since 2001, a number roughly the size of a single residency program's alumni, against a workforce of hundreds of thousands of surgeons.
The three people who could have unblocked the surgeon at her kitchen table, an engineer, a peer who had already cleared a device at her target CDRH office, and a licensing officer who would tell her the truth about her contract, all exist somewhere in American medicine tonight. None of them know she is looking.
Part of a series on the missing professional infrastructure of healthcare. Previously: The Invisible Negative Result
Evidence note: the core patenting figures (275,260 surgery-related patents, 9,008 held by surgeons, 2,164 individuals, 97% male, 462% growth) are from a 2022 Annals of Surgery bibliometric study using USPTO-to-ACS Fellow name matching, a methodology that can undercount surgeons whose names did not match cleanly and should be read as a careful estimate rather than an exact census. The 510(k) refuse-to-accept rates and the $31 million to $94 million, three-to-ten-year development estimate are from a 2023 Journal of Clinical and Translational Science analysis; per-office rejection rates as high as 60% describe specific CDRH offices, not the pathway overall. Stanford Biodesign's fellow and student counts are self-published by the program. This article does not constitute legal, regulatory or intellectual-property advice; readers navigating an actual invention disclosure should consult their institution's technology transfer office and independent counsel regarding ownership, royalty and compliance questions specific to their employment situation.
