There is a nurse on a medical ICU somewhere who has worked that unit for twenty-two years. She can walk past a room and know that the patient in it is going to deteriorate in the next four hours, before anything shows up in the vitals. Ask her how and she will say something unsatisfying about colour, or breathing, or a look. She is right often enough that the residents have learned to take her seriously.
She is retiring in March.
The hospital will conduct an exit interview. It will ask why she is leaving, whether she felt supported, and whether she would recommend the organization as a place to work.
It will not ask the only question that matters: what do you know that nobody else here knows?
On her last day, she will hand over her badge. Twenty-two years of pattern recognition will walk out of the building, and the institution's formal record of it will be an entry in a human resources system marking her employment as ended.
This happens continuously, everywhere, and healthcare has no mechanism to notice.
The loss curve is steepening
The demographics make this the wrong decade to keep ignoring the problem.
Physicians. According to AAMC workforce analysis, physicians aged 65 and over made up 17 percent of the active workforce in 2021, with those aged 55 to 64 accounting for another 25 percent. That means more than a third of currently active physicians are on track to retire within a decade, against a projected shortage of 13,500 to 86,000 physicians by 2036.
Nurses. NSI's 2026 report puts RN turnover at 17.6 percent, up 1.2 points year over year, with retirement now the third most common cause of voluntary RN resignation. Average bedside RN tenure is 6.8 years. And revealingly: only 62.4 percent of hospitals report a retention strategy targeting tenured nurses specifically, versus 80.8 percent for newly hired nurses.
Read that last comparison again. The population carrying the most irreplaceable knowledge receives the least targeted retention attention. NSI itself flags the retirement wave as an unaddressed knowledge-loss risk.
A systematic review in the Journal of Health Organization and Management concluded that hospital doctor turnover creates "substantial intangible costs in terms of the loss of organisational and tacit knowledge."
Intangible, in this context, means unmeasured. It does not mean small.
What is actually being lost
"Institutional knowledge" is too vague to act on. Be specific about what disappears.
Pattern recognition that precedes the data. The nurse who knows the patient is crashing before the monitor does. This is real, it is the product of thousands of observations, and it exists in no document.
The reason the protocol has that step. Every mature clinical pathway contains a step that looks unnecessary and exists because of something that happened in 2014. When the person who remembers leaves, the step becomes an unexplained inefficiency, and eventually someone removes it.
The failure library. The drug interaction that keeps getting missed. The transfer that always goes wrong on weekends. The equipment that fails in a specific way. This is negative knowledge, and negative knowledge is never documented anywhere.
The relationship map. Who to call at the other hospital. Which specialist actually takes these patients. Who to ask when the system says no. This is the trust graph, held individually, and it evaporates completely on departure.
And in shrinking subspecialties, the exposure itself. In fields where training pipelines are contracting, a retiring physician may be one of a small number of people who has managed a given condition at volume. That is not organizational knowledge. That is clinical knowledge that will not exist anywhere afterward.
Meanwhile the institutional response is oriented entirely around documents. Protocols, order sets, charts, policies. Healthcare is genuinely excellent at preserving artifacts and has no mechanism at all for preserving the judgment that makes the artifacts mean anything.
The second failure: they want to come back, and cannot
Here is where the story turns from loss to something closer to absurdity.
The retiring workforce is not, for the most part, leaving because it wants to stop contributing.
- Roughly 160,000 US physicians are inactive, with about 10,000 a year becoming eligible to re-enter.
- Research indicates 78.9 percent of inactive physicians are open to re-entry.
- Around 37 percent of retired physicians say they want to volunteer.
- Survey work found the top barriers to be limited part-time or flexible opportunities (42.5 percent) and limited retraining (40.4 percent), with flexible work (51.5 percent) and desire to volunteer (39.6 percent) as leading motivations.
- Retired physicians consistently report that their primary fears about retirement concern loss of purpose and social interaction rather than money.
- About 19 percent of locum physicians began locum work after retiring.
So: a large, willing, deeply experienced population, whose stated barriers are coordination failures rather than competence or interest.
Now look at what re-entry actually requires.
Formal re-entry assessment programs cost between roughly $6,750 and $20,000 or more, with structured programs running around eight weeks. State requirements are a patchwork; boards typically require formal re-entry after roughly 2.8 years out of practice, with a range from one to ten years, and close to half of boards have no formal re-entry policy at all.
And then there is the step that actually stops people. In the words of physician-facing guidance on returning to practice:
> "One of the hardest things for physicians re-entering clinical medicine can be having to find someone that is willing to serve as their proctoring or monitoring physician."
The binding constraint on physician re-entry is not competence assessment. It is sponsorship.
A program can certify readiness in eight weeks. The year is lost trying to find a working physician willing to vouch, supervise, and host. That is a matching problem, not a regulatory one, and nobody brokers it. Programs assess. Boards license. Neither places anybody.
The third failure: identity switches off at exactly the wrong moment
There is a final structural irony that makes all of this harder than it needs to be.
On the day a physician retires, several things happen simultaneously. The institutional email is deactivated. The directory listing disappears. Hospital privileges lapse. Society membership frequently lapses with them.
Every piece of verified-identity infrastructure that let anyone confirm who this person was and what they were qualified to do is switched off at once.
That would be a minor inconvenience if retirement were a clean exit. But as the numbers above show, for a large share of physicians it is a transition into a different, lower-intensity mode of continued contribution: volunteering, locums, teaching, mentoring, or eventual re-entry.
And regulators are moving in the opposite direction. The emergency medicine literature notes that where re-entry once required little beyond resuming work, "regulatory agencies are now increasingly requiring additional training, standardized testing, and fitness to practice evaluations prior to restarting clinical work."
So the requirements to prove yourself are tightening at precisely the moment your ability to prove yourself has been dismantled.
The reason is structural rather than malicious. Every verification system in healthcare is anchored to an active employment or privileging relationship. Credentialing files, privileging records, payer enrollment: all assume a current institutional sponsor.
Retirement is the single most predictable and universal identity-continuity event in a professional career, and it is exactly the case the infrastructure was never designed to handle.
Healthcare's identity systems are built around "currently employed" rather than "verified," which is why they fail this population predictably and by design.
The reframe: the retirement cliff and the shortage are the same population
Here is the observation that ought to reorganize how workforce policy treats this.
Healthcare talks about two separate crises. A physician shortage projected at up to 86,000 by 2036. And a retirement cliff, with more than a third of the workforce departing within a decade.
These are the same people, counted twice, on opposite sides of the ledger.
Roughly 160,000 inactive physicians, of whom nearly four fifths say they would return in some capacity, is one of the largest available reserves of trained clinical capacity in the country. It requires no new medical schools, no new residency slots, and no fifteen-year pipeline.
It requires solving three coordination problems: flexible opportunities, a route back, and someone willing to sponsor.
Every one of those is a matching problem. None is a competence problem. And the pandemic demonstrated that fast paths are possible when the will exists: 29 states issued expedited licensure by April 2020.
What would actually work
The knowledge exit interview. Not the HR one. A structured, recorded, ninety-minute conversation with anyone leaving after long tenure, covering: what do you watch for that is not in the protocol, what fails around here and why, what do you wish you had known in year one, and who do you call outside this building. Ninety minutes, indexed, kept. Almost no organization does this and every organization should. It is the cheapest knowledge-retention intervention available.
Identity that survives the badge. A verified professional record owned by the individual rather than the institution, which does not switch off on the last day. This is the structural fix, and its most valuable property is that it makes the re-entry review faster and cheaper for exactly the people who need it most.
A proctor exchange. The single highest-leverage intervention in physician re-entry. If the binding constraint is finding a working physician willing to supervise, then a mechanism that matches returning physicians to willing proctors, including former co-residents and former trainees, unblocks a population of tens of thousands.
Flexible contribution roles, designed as such. The stated barrier is the absence of flexible options, cited by roughly 42.5 percent. Answering peer questions, teaching, proctoring, second-opinion review, mentoring new attendings, and covering short gaps are all valuable, all part-time, and all currently arranged by luck.
And a specific recognition: retired clinicians are the safest listeners in medicine. They sit outside every peer review chain, every promotion committee, and every institutional interest. For confidential peer support, for the physician who has just been sued or investigated, for the new attending who does not know who to ask, that independence is not incidental. It is the whole thing.
What you can do now
If you are approaching retirement
Write your own knowledge document. What you watch for, what fails, what you would tell a new colleague, who you would call for what. Nobody will ask you for it. It is the most valuable thing you will produce in your final year of practice.
Decide what continued contribution looks like before you stop. The evidence is consistent that the losses people regret are purpose and connection rather than income. Deciding in advance whether you want to teach, proctor, answer questions, or do occasional clinical work is far easier than reconstructing a route back afterward.
Keep your credentials alive if there is any chance you will return. Boards typically require formal re-entry after roughly 2.8 years out. Understanding your own state's threshold before you cross it can save you $6,750 to $20,000 and a year.
Say explicitly that you are available. To your former trainees, your department, your specialty society. The single sentence "I am happy to be called about X" is what converts a retired expert into a reachable one, and almost nobody says it.
If you are trying to return
Find your proctor first. It is the binding constraint, and everything else in the process is more tractable. Start with people who trained with you or who you trained, because that relationship carries the trust the arrangement requires.
Check your state board's actual policy. Requirements vary enormously and close to half of boards have no formal re-entry policy, which cuts both ways: less structure can mean fewer obstacles or more uncertainty.
Consider structured programs realistically. They cost $6,750 to $20,000 or more and run around eight weeks. That is a real expense and it is not the part that usually stops people.
If you lead a department or a hospital
Institute the knowledge exit interview this quarter. Ninety minutes, structured, recorded, for every long-tenured departure. It is close to free and it is the only chance you get.
Look at the retention gap. With 62.4 percent of hospitals targeting tenured nurse retention versus 80.8 percent targeting new hires, your organization is probably underinvesting in the group whose departure costs you most.
Create genuinely flexible roles. The stated barrier is the absence of them. A half-day teaching commitment, a proctoring arrangement, or a defined peer-answering role is not difficult to construct and is not on offer anywhere.
Sponsor a re-entry. If you have a workforce gap and there is an experienced physician nearby who needs a proctor, you are looking at each other across a coordination failure that one phone call would resolve.
Frequently asked questions
How many physicians are approaching retirement? AAMC analysis found physicians aged 65 and over comprised 17 percent of the active workforce in 2021, with another 25 percent aged 55 to 64, meaning more than a third of active physicians are within roughly a decade of retirement, against a projected shortage of 13,500 to 86,000 physicians by 2036.
Do retired physicians want to keep working? A large majority express interest in some form of continued contribution. Research indicates 78.9 percent of inactive physicians are open to re-entry and roughly 37 percent of retired physicians want to volunteer, with the leading stated barriers being limited flexible opportunities (42.5 percent) and limited retraining (40.4 percent) rather than lack of interest.
What does physician re-entry to practice require? It varies by state, with boards typically requiring formal re-entry after roughly 2.8 years out of practice, though the range runs from one to ten years and close to half of boards have no formal policy. Structured assessment programs cost roughly $6,750 to $20,000 or more and run around eight weeks. The most commonly cited practical obstacle is finding a practising physician willing to serve as proctor or monitor.
Why is finding a proctor so difficult? Because no mechanism exists to match returning physicians with willing supervisors. Assessment programs evaluate competence and state boards issue licenses, but neither places anyone. The arrangement depends on a personal favor from a working physician, and there is no broker.
What knowledge is lost when experienced clinicians leave? Primarily tacit knowledge: pattern recognition that precedes measurable changes, the reasons behind protocol steps, the library of local failure modes, and the relationship map of who to call elsewhere. A systematic review of hospital doctor turnover describes substantial intangible costs from loss of organisational and tacit knowledge. Standard exit interviews ask about reasons for leaving rather than about knowledge.
What happens to a physician's professional identity at retirement? Most of it switches off simultaneously: institutional email, directory listings, privileges, and frequently society membership. Because credentialing infrastructure is anchored to active employment or privileging relationships, there is no mechanism to maintain verified identity through a gap, which makes any later return harder and more expensive precisely when regulators are tightening re-entry requirements.
The bottom line
More than a third of America's physicians are within a decade of retiring, into a projected shortage of up to 86,000. Nursing is losing tenured staff at 17.6 percent turnover with retirement now the third leading cause, while hospitals target retention efforts at new hires far more than at the experienced nurses whose departure costs the most.
What leaves with them is the part of clinical practice that never got written down: the pattern that precedes the vital signs, the reason for step three, the list of things that reliably fail, and the phone numbers of people who will help.
And this is the strange part. Most of them do not want to disappear. Nearly four fifths of inactive physicians say they would come back in some form. What they say stops them is the absence of flexible options and the difficulty of finding one working colleague willing to sponsor their return.
We are simultaneously running a workforce shortage and a knowledge loss crisis, and both involve the same people, who are standing right there, saying they would like to help.
On her last day, that ICU nurse will hand back her badge, and the system will record that a full-time equivalent has been vacated.
Nobody will ask her what she knows.
Part of a series on the missing professional infrastructure of healthcare. Previously: Rare in Childhood, Invisible in Adulthood
Evidence note: workforce age distribution and shortage projections come from AAMC's "The Complexities of Physician Supply and Demand" (2024). Nursing turnover, tenure, and retention strategy figures come from NSI Nursing Solutions' 2026 National Health Care Retention and RN Staffing Report. Tacit knowledge loss is discussed in Seathu Raman, McDonnell and Beck, Journal of Health Organization and Management (2024). Re-entry figures come from AAMC reporting, Annals of Internal Medicine survey research, CPEP program information, and physician-facing practice guidance. Regulatory tightening of re-entry requirements is described in Guth et al., Journal of the American College of Emergency Physicians Open (2020). Some figures on retired physician volunteering intentions and locum entry after retirement come from survey and industry sources that could not be independently re-verified and should be confirmed before republication.
