The Endless Residency

Why Surgical Training Never Really Ends

1. A Simple Procedure

Early in my residency, I placed a suprapubic tube. It is a routine procedure, the sort every urologist is expected to do safely. A small incision is made above the pubic bone, a trocar is advanced into the distended bladder, and a catheter is passed. An attending appropriately supervised the case. I was neither unsupported nor unprepared.

The patient died.

The trocar created a small peritoneal defect. Urine leaked intraperitoneally, the ileus results, and the patient vomited and aspirated. A routine procedure, performed acceptably, killed someone.

I reviewed the case obsessively. I could not find the exact thing I had done wrong. The usual safeguards had been in place. The bladder was distended, the landmarks were correct, and there had been no bowel injury. The absence of a clear error was, in a way, worse. What remained was harder to name than guilt: the possibility that I was morally responsible without being able to identify the precise error. I sensed that error must exist, because the patient was dead and I had been the one who did the procedure.

I came very close to leaving surgery. Not because anyone told me to, but because I could not figure out whether staying was the ethical thing to do.

What I did not understand then was the distinction between culpable error and stochastic harm. Training had given me no framework for separating the two. In a wicked learning environment, feedback is noisy, delayed, and often misleading: good decisions can end badly, and bad decisions can look fine for a while.¹ A complication after a correctly performed procedure is not evidence of incompetence; it is evidence of variance.

But the training culture offers no tools for distinguishing between the two. You are encouraged to feel bad, because feeling bad looks like accountability. M&M can review a complication in technical detail while still leaving untouched the deeper question of how a surgeon is supposed to metabolize uncertainty, variance, and grief.

I stayed. I eventually developed a high-volume complex reconstructive practice. I participated in first-in-field robotic procedures. I trained residents and fellows. Whatever contribution I have made would not exist if I had left residency during those weeks when leaving felt like the most therapeutic option. I do not know how many people in similar situations did leave, or how many chose career paths designed to minimize the number and complexity of cases they would perform. I suspect the system treats such departures as failures of resilience rather than predictable outputs of a structure that provides no framework for processing grief.

Surgery requires you to do things to people that carry real risk of killing them, including in situations where you perform the procedure correctly and the outcome is still a death. The training model requires that patients who are, in some meaningful sense, teaching cases bear this risk.

We obtain their consent. We tell them a trainee may be involved. What that conversation actually sounds like is something like: I will be performing your procedure. I will have a resident assisting me, and they may participate in portions of the case under my supervision. This is standard for an academic medical center. Do you have any questions?

The patient, lying in a hospital gown thirty minutes before surgery, typically says no.

What they cannot fully know, and what I cannot fully quantify, is that my own sense of proficiency becomes less reliable when I am operating on a new learning curve. No consent form captures this, because the surgeon does not fully know it either. They are not being deceived. But they are bearing a cost the system prefers not to name.

2. The Surgeon Who Built the System

In the 1880s, William Stewart Halsted was among the most technically gifted surgeons in New York. He had trained in Vienna, Leipzig, and Würzburg, and returned to Roosevelt Hospital and Bellevue with a combination of radical ambition and obsessive precision that his contemporaries found unusual. He performed one of the earliest documented human-to-human blood transfusions on his own sister, drawing blood from his arm to save her from postpartum hemorrhage in Albany in 1881. When his mother developed cholecystitis, he operated on her on the kitchen table. Both survived. These episodes sketch the man: rapid improvisation under extremis, comfort with invasive risk, and the personal risk tolerance of someone free-soloing a rock face.²

In the mid-1880s, he began experimenting aggressively with cocaine as a local anesthetic on himself, pushing peripheral nerve blocks that required precise dissection in relatively awake patients. The work was genuinely innovative. He also became addicted. His colleagues arranged treatment at Butler Sanatorium in Providence, where they substituted morphine for cocaine. It managed the acute crisis and created a chronic one. His New York career was over.

William Welch brought him to Baltimore, where Halsted spent the remaining thirty-six years of his life managing a morphine addiction that was an open secret among his colleagues. From that position, he designed the residency system that every surgical specialty still uses today. Residents literally resided in the hospital. Training lasted seven or more years. The chief resident functioned as a feudal lord over the service. The implicit promise was direct: endure the pyramid and you would emerge a complete surgeon.

As a production engine for surgical talent, the system was extraordinary. Harvey Cushing founded modern neurosurgery out of Halsted’s program. Hugh Hampton Young established what is recognized as the first formal urology residency and performed the first radical perineal prostatectomy for cancer in 1904. Nearly every academic urology program in the country traces its intellectual lineage through Young to Halsted.

I still use Young’s dissection approach when I operate on the prostate via the perineal route. A technique developed in the early 1900s, by a man who learned his craft from a morphine-dependent surgeon, remains current enough to be taught and used today. That says something about both Young’s innovation and the uneven pace of change in surgery.

The technical principles Halsted codified (meticulous hemostasis, gentle tissue handling, respect for anatomy) remain current. But so does something else he installed, something less often named: a hidden curriculum that instructs surgeons to project strength, conceal fragility, and absorb the cost of the work in silence and alone. We have added wellness committees and duty-hour regulations and required lectures on burnout, but Halsted’s emotional architecture is intact.

Halsted’s model rested on an assumption that no longer holds: that training could be completed once, and that the resulting competence would remain durable for decades. Before him, surgical training in the United States was pure apprenticeship: you attached yourself to a surgeon, watched, and eventually left to practice. No standardized curriculum. No graduated responsibility. Quality depended entirely on whom you had access to.³

His residency was a genuine improvement. It made training legible, reproducible, and substantially safer. But it required surgical knowledge to change slowly enough, and to be organized narrowly enough, that seven years of immersion could produce a surgeon whose core technical repertoire would remain serviceable for decades.

That assumption was reasonable then. It is not reasonable now. The training itself has contracted while the knowledge has expanded. Halsted’s residents trained for seven years or more with no duty-hour restrictions. Today’s residents train for fewer total hours while trying to master a field that is broader, faster-moving, and less stable than the one Halsted imagined.⁴ The 80-hour workweek improved resident well-being in measurable ways, though evidence on patient safety outcomes remains mixed. The net effect is a structural mismatch: a finite-training model applied to a field that no longer stabilizes.

Residency offers pockets of kind feedback, but it is preparing surgeons for a career that is fundamentally wicked, and the handoff happens almost without ceremony. You finish residency on a Friday. On Monday you are the attending. No one really marks the shift. One day you are supervised; the next you are the one expected to know.

3. The Red Queen and the Radiated Vein

I began my career operating under Halsted’s assumption of a stable plateau: I trained in open surgery, learning to tie knots, retract tissue, and remove prostates by hand. Laparoscopic surgery forced me to relearn how to operate through a screen, watching a two-dimensional image with instruments that pivoted on a fulcrum. Endoscopic surgery required a different set of spatial assumptions entirely.

During my six years of residency, multi-port robotic surgery replaced much of what laparoscopy had recently established, and I spent months rebuilding motor patterns. Then single-port robotic surgery compressed the workspace further, and changed the instrument geometry again. At each transition, procedures I had been performing confidently became either obsolete or unrecognizably different. Almost everything I did routinely as a resident and fellow I have substantially modified or abandoned.

In Through the Looking-Glass, the Red Queen tells Alice that it takes all the running you can do just to stay in the same place. That is the experience of practicing surgery today. The feeling of being a trainee, that specific sensation of attempting a motion and being slightly off in a way you cannot yet diagnose, never fully resolves. It merely becomes less socially acceptable to admit.

Not long ago I was deep in a single-port robotic case, working in the narrow pelvis of a patient who had already undergone two previous operations. Scar tissue obliterated the anatomy. Fibrosis encased the ureter, tethering it to structures it should not have been touching. Without the robotic platform, few surgeons would have attempted this case.

I had dissected the ureter and was suturing the anastomosis when the tip of my needle brushed against the radiated internal iliac vein. What happened next was not cinematic; it was hydraulic. A one-centimeter opening in a vein stented open by radiation and fibrosis. The field filled with blood. The 10x magnified 3-D view on the console became a red screen. Massive transfusion protocol. Coolers of blood in transit. Attending anesthesiologists rushing in. People who were relaxed were now moving quickly and not talking unless they had something important to say.

I got control. Pressure with one instrument, suction with another, and precise suturing within a field that did not want to be sutured.

At M&M the next month, that is the short version. The patient recovered; the reconstruction worked. Technically, the case was a success.

But what lives in my memory is the interval between the injury and the control, which was probably four minutes and felt like months. And what followed: the familiar post-complication inventory, the review of whether the approach was right, whether different movements would have kept me out of that vein. There is no clean answer. The anatomy was hostile. The platform was correct. The injury was a consequence of operating where the margin between a good reconstruction and a vascular catastrophe is measured in millimeters.

I was not a junior resident. I was a high-volume attending at the frontier of robotic reconstruction, performing a case few surgeons would attempt. And in the moment of that vascular injury, I experienced novice-like uncertainty, because platform change resets parts of expertise. The platform allowed me to attempt something few would, and I experienced an edge case that I could not prepare for.

Halsted’s model assumes that expertise, once built, is stable. Technological change repeatedly sends even experienced surgeons back onto learning curves that the credentialing system barely acknowledges.

4. The Prisoner’s Dilemma of Teaching

When the adrenaline of a vascular crisis fades and I step away to write the operative note, the electronic medical record asks for nothing about the psychological weight of what just happened. It simply demands the correct billing code. This abrupt shift from the visceral to the administrative highlights a broader failure in how modern medicine functions: the system is quietly eroding real training while loudly measuring everything except it. The hospital generates revenue from residents seeing patients and assisting cases, but it operates on a currency that inherently punishes the act of instruction. Teaching consumes attending time that has a measurable opportunity cost in Relative Value Units—the metric by which US physicians are compensated under the Medicare fee schedule. Time spent demonstrating a tissue plane or watching slow suture placement generates nothing billable.

The attending who spends 45 extra minutes letting a resident work through a difficult anastomosis is economically penalized relative to the one who takes over and finishes. For an anastomosis that leaks because the resident was learning, the penalties extend well beyond economics. We protect time for writing papers that nobody reads but not for teaching residents how to operate. That tells you everything about what academic medicine values.

As a resident, my program tasked me with selecting the recipient of a teaching award. I gave it to the attending who was most rarely present during my cases. The one who had taught me, in effect, by absence. We called it a teaching award in absentia. I meant it partly as a joke. Looking back, it was a precise description of what the incentive structure had produced.

Now zoom out. The resident operates under work-hour restrictions that have compressed training time, while documentation burden consumes a growing share of what remains. The resident’s rational strategy is to optimize for what is measured: board passage, recommendations, credentialling, and the next job. The patient rationally prefers the attending to do the case. The attending rationally wants anastomoses not to leak and to come home earlier to family.

Every individual making the individually rational choice produces a collective outcome that nobody wants: a generation of surgeons less well-prepared than they could be. Nobody defected maliciously. The incentive structure produced the outcome. This is the prisoner’s dilemma of surgical education, and it has no individual solution.

5. The Wicked Classroom

The knowledge expansion means the career never stabilizes. Each new platform, each new technique resets you briefly to the position of a beginner. The benefit to the patient accrued through any learning curve has to be re-earned continuously, by every surgeon, in every new domain they enter, for the rest of their career.

I have written before on how I thought about postoperative fever, drawing on a framework absorbed in residency: fever on postoperative day one was probably atelectasis. Attending surgeons repeated it so often and so confidently during my training that it felt like physiology rather than dogma. Yet there is no evidence supporting the association.⁵ I knew it abstractly, but not in the way that changes practice.

This is how mistakes propagate. Not dramatic therapeutic misadventure, but a piece of received knowledge, transmitted with confidence, that turned out to be wrong. The question is not whether this happens (it always happens) but whether you have built the feedback structures that let you find out before you have propagated the error through another generation of trainees. In a wicked environment, you can teach the wrong lesson for twenty years and never know it.

Burnout often lives in the gap between what the role requires you to project and what the work actually feels like from the inside. The entire social architecture of surgery is built around the premise that the experienced surgeon has moved past the learning curve and is operating on a stable plateau. The honest internal experience is: I am still learning how to do surgery. I will always be learning how to do surgery.

When a profession demands certainty that reality cannot supply, people spend enormous energy pretending the mismatch is not there. That energy comes from somewhere. Usually, it draws from the reserve that would otherwise go toward learning, honest self-assessment, and the practices that produce genuine improvement. Halsted managed this gap chemically. We have replaced that with more respectable but not necessarily more honest coping mechanisms.

    *

Surgery remains a glorious field that offers something almost nothing else does. You can intervene directly in another person’s suffering, with your insight, your hands, your skills, right now. The privilege of that is not diminished by broken incentive structures.

The patient on the table is real regardless of what the RVU calculation says. That reality is worth protecting, which is exactly why the structural failures matter as much as they do. The only real defense is operational humility: review the case, track the outcomes, change course when the evidence demands it, and ask the questions that make you uncomfortable. Distinguish pain from guilt. You will have complications. Many will reflect genuine errors. Many will not. The ability to tell the difference is the core skill that no residency explicitly teaches.

The residents I train now will practice in a world I cannot fully imagine. The best thing I can offer them is not the performance of mastery, but a model of what it looks like to keep learning as the map changes.

I carry that with me as I sit at the console for the thousandth time. I slide my bare fingers into the controllers, accepting that the scarred tissue will inevitably surprise me. I accept, too, the terrifying, reasonable possibility that in the days ahead, I might feel exactly like that young resident again: a humbled man forced to tell a patient’s children how my hands led to her death.

Residency teaches you to operate. What follows teaches you that residency never ends.

Notes

1. Robin Hogarth, Educating Intuition (University of Chicago Press, 2001). Extended in Hogarth, Lejarraga, and Soyer, “Lessons From Experience,” Current Directions in Psychological Science (2015).

2. Gerald Imber, Genius on the Edge: The Bizarre Double Life of Dr. William Stewart Halsted (Kaplan, 2010). David Oshinsky, Bellevue: Three Centuries of Medicine and Mayhem at America’s Most Storied Hospital (Doubleday, 2016). On Halsted’s residency structure and institutional spread, see Nick Carter’s centenary survey (1952), discussed in Bhatt and Rao, “Necessity Is the Mother of Invention,” Canadian Journal of Surgery (2021).

3. Oshinsky, Bellevue, op. cit. For amputation mortality in the pre-antiseptic era: Ira Rutkow, Bleeding Blue and Gray: Civil War Surgery and the Evolution of American Medicine (Random House, 2005).

4. ACGME implemented 80-hour workweek limits in 2003, with further restrictions in 2011. The FIRST trial (New England Journal of Medicine, 2016) and iCOMPARE trial (New England Journal of Medicine, 2019) found non-inferior patient outcomes under flexible vs. standard duty-hour rules. A 2014 Annals of Surgery systematic review found no overall improvement in patient outcomes, with concern about complications in high-acuity surgical patients. Evidence on resident safety hazards (motor vehicle crashes, needlestick injuries) more consistently favors restrictions. See PSNet/AHRQ for synthesis.

5. Michael N. Mavros, George C. Velmahos, and Matthew E. Falagas, “Atelectasis as a Cause of Postoperative Fever: Where Is the Clinical Evidence?” Chest 140, no. 2 (2011): 418–24. No clinical evidence supporting the atelectasis-fever association. Multiple subsequent publications confirm the myth persists in teaching despite absence of supporting evidence; see Abdelmaseeh et al., “Debunking a Mythology,” Clinical Imaging (2024).