A surgeon fifteen years into practice is filling out a privileging application at a new hospital, and it is asking her for something she assumed she already owned: her own case volumes.
She trained at one institution, has since worked at three others, and every one of those employers kept its own version of her procedural history, in its own system, for its own accreditation and liability purposes. None of them handed her a copy on the way out the door, because none of them were built to. She spends the better part of two weeks emailing former department administrators, most of whom have moved on themselves, trying to reconstruct fifteen years of case counts one hospital at a time, so a credentialing committee can decide whether she is allowed to keep doing the job she has already been doing for a decade and a half.
Meanwhile, a pilot with the same fifteen years of experience opens a logbook that has been hers since flight school. Every hour, every aircraft type, every instructor's signature, all in one continuous record that has traveled with her through every airline, every job change, every certification renewal, because aviation decided a century ago that a pilot's flight record belongs to the pilot.
Medicine never made that decision. A surgeon's procedure history, a clinician-educator's teaching record, a physician's entire longitudinal account of what they actually did, belongs, in practice, to whichever employer happened to be watching at the time, and it is orphaned completely at every single job change.
The record is unreliable even before it gets orphaned
Before this article gets to what happens to the record after someone leaves, it is worth being honest about how shaky the record already is while someone is still there.
Only 11.3 percent of surgical residents describe their own case logs as highly accurate. 51.4 percent report backlogs of three months or more in entering their cases. 68.8 percent find the system burdensome, and 64.2 percent find it difficult to use, according to a 2016 multispecialty survey by Cadish and colleagues in the Journal of Surgical Education.
The coding accuracy underneath those self-reports is worse than residents themselves realize. General surgery residents' case-log coding accuracy measured 52.82 percent, ranging from 43 to 60 percent depending on the specific study population, even though 74 percent of residents in that same research believed they were coding accurately (Balla et al., Journal of Surgical Education, 2016). A broader 2020 study by Collins and colleagues found overall logging accuracy of 72 percent, with wide variation by specialty: 91 percent in orthopedics, 69 percent in general surgery, and just 58 percent in neurosurgery.
Read those numbers together. The confidence residents have in their own case logs is not just modest, it is actively miscalibrated: three-quarters believe they are logging accurately while barely half actually are. This is the foundational record that every subsequent credentialing, privileging, and competence decision in a proceduralist's career gets built on top of.
And then it is orphaned entirely
Here is where the problem moves from "unreliable" to "structurally abandoned," because the unreliable record does not even survive the transition it was theoretically built to serve.
The ACGME Case Log System captures a resident's cases during training, with the program director attesting to the entries. After graduation, whether a physician retains meaningful access to their own logged history varies by specialty and is not published anywhere in accessible form. What happens after that is a fragmented, institution-by-institution patchwork: attendings rely on EHR reports pulled at whichever hospital they currently work, personal spreadsheets kept privately and inconsistently, and case lists compiled separately for board certification through the relevant specialty board. Clinician-educators face the same problem in a different domain: teaching hours and mentorship records live inside systems like MedHub or New Innovations, owned outright by the GME office at whatever institution the educator happens to work for at the time, with no export path when they leave.
None of it is portable. None of it survives a job change intact. A mid-career interventionalist with case volumes spread across three hospitals' separate EHR systems cannot produce a single, continuous account of her own procedural history, because no such account exists anywhere except in fragments she does not control.
What this actually costs, in time and in delayed careers
The abstraction of "the record is fragmented" becomes concrete very quickly once you attach a number to how much time it costs to reconstruct.
Roughly 40,000 residents and fellows graduate each year, and an estimated 50,000 physicians change jobs annually. Each one of them plausibly spends 5 to 20 hours reconstructing procedural volumes for a new privileging process, according to the underlying research estimate. At a conservative $150 per hour of physician opportunity cost, that works out to $70 million to $280 million a year in clinician time alone, before counting the medical staff services time spent verifying those reconstructed numbers, and before counting the 112-day average gap between signing a contract and actually starting work, a delay to which incomplete or slow credentialing verification is a known contributor.
That is not a one-time inconvenience. It is a recurring tax, paid by essentially every proceduralist in the country, every single time they change employers, for the entirety of a career that can span three or four decades and multiple institutions.
The deeper loss: the attestation, not just the count
There is a more subtle failure buried inside this problem, and it matters more than the raw hours lost to reconstruction, because it touches the actual substance of what a case record is supposed to prove.
A case record is not really a number. It is a graph: a trainee, a supervising surgeon, a co-surgeon, an institution, a specific procedure, on a specific date. The value in that structure is almost entirely in the attested edges, the fact that a specific named supervising surgeon watched a specific trainee perform a specific procedure and is willing to vouch for what they saw. Today, that structure gets flattened into a bare CPT code count with no counterparty attached at all. The supervising surgeon's actual attestation, the real competence signal buried inside the transaction, gets discarded the moment the case gets logged as a number rather than a verified, signed claim.
This is precisely the aviation counterexample again. A pilot's logbook entry is not just "flew 2.3 hours in a Cessna 172." It is that claim, signed by a specific instructor or captain who was physically present and is putting their own name behind the record. That signature is what makes the logbook trustworthy years later, to a different employer who was never in the room. Medicine's case logs strip that signature out entirely, reducing a rich, verifiable claim about supervised competence down to an unverifiable, easily miscoded count.
Why nobody owns this
Run through every plausible owner of this problem, and the same pattern that recurs throughout this series shows up again: each institution's remit ends exactly where the physician's need for a continuous record begins.
ACGME's remit ends at graduation. It was built to accredit training programs and verify that residents meet case-volume requirements during training. It has no mandate, and no obvious funding path, to maintain a lifelong post-graduation record for someone who is no longer a trainee in any program it accredits.
Specialty boards keep case lists only long enough to establish exam eligibility, not as an ongoing career record. Once a physician clears the relevant certification hurdle, the board's interest in that specific data effectively ends.
EHR vendors serve the hospital that pays them, not the individual clinician who generates the data inside that system. A physician's procedural history inside an EHR is, functionally, the hospital's data, not hers, regardless of the fact that she is the one who performed the procedures.
Hospitals have no institutional incentive to help a departing physician assemble a portable record. Once someone has resigned, the hospital's remaining interest in that person's historical data is limited to its own liability and accreditation needs, not the departing clinician's next job.
Medical societies with registries, such as the American College of Surgeons' NSQIP, are built around institutional outcomes, not individual portable careers. NSQIP is a genuinely valuable dataset for measuring hospital-level quality. It was never designed to hand an individual surgeon a continuous, personal, portable account of her own case history across every hospital she has worked at.
Doximity has no procedural data at all. Its entire product is built around messaging, visibility, and networking, not the underlying clinical-activity record this problem actually requires.
Every party touching this data is optimized for institutional accreditation or liability, not for the individual whose entire career depends on being able to prove what they actually did.
Why this is becoming more urgent, not less
Three forces happening right now are raising the cost of leaving this gap unaddressed.
Competency-based medical education, built around entrustable professional activities, is increasing demand for longitudinal procedural evidence specifically, not just a raw case count, which means the shift toward richer, attestation-based competence evidence is already underway on the training side, with nothing built yet to carry that richer evidence forward past graduation.
Focused-practice designations, which certify a physician's specific expertise within a broader specialty, depend entirely on demonstrable volume history that, as described above, most physicians cannot actually produce on demand.
AI-generated operative notes are making procedural volumes far more machine-readable than they have ever been, which is a genuine opportunity: the raw material for an accurate, low-friction personal record is becoming easier to capture automatically than at any point in the past. ACGME's own move toward a centralized "ACGME Cloud" data platform in 2025-2026 signals real momentum toward data centralization. It remains, notably, institution-side rather than physician-side: a better version of the same architecture that ends at graduation, not a different architecture that follows the individual.
What would actually work
A signed logbook entry, not a bare count. Every entry should capture the procedure or teaching event, the date, the physician's specific role, the counterparty who can vouch for it, and that counterparty's actual attestation signature, restoring the structure that a raw CPT-code tally currently discards.
Owned by the individual, exportable in privileging formats, importable from existing systems. The record needs to live with the physician across every job change, not inside any single employer's system, while remaining compatible with the ACGME Case Log CSV export and the formats credentialing committees already expect to receive.
Countersignature from a specific, named, reachable peer. The value of a case record is almost entirely in the attestation. A system that lets a supervising surgeon sign off on a case once, permanently, in a form that travels with the trainee for the rest of their career, restores the aviation-logbook model this profession never adopted.
Built for teaching and mentorship, not only procedures. A clinician-educator's contribution, teaching sessions, mentorship hours, answered peer consults, deserves the same portable, attested treatment as a surgical case count, since promotion and academic advancement depend on exactly this kind of longitudinal record and it currently lives, unportably, inside a GME office's internal spreadsheet.
Strict guardrails on patient identifiers and peer-review confidentiality. Any credible version of this has to capture volumes and attestations only, never patient-identifying information, and has to be built with explicit attention to state peer-review confidentiality law and the discoverability risk that legal counsel would reasonably flag around any self-kept clinical log.
Institutional acceptance built in from the start, not bolted on afterward. A portable record only has value once privileging committees are willing to treat a peer-countersigned external log as legitimate supporting evidence, which means the design work here is as much about building institutional trust and format compatibility as it is about the underlying data structure.
A natural role for retired physicians as countersigners. A retired surgeon has every reason to be a generous, reliable countersigner for the historical record of trainees they personally supervised years or decades earlier, and currently has no structured way to do so once they have left clinical practice themselves.
What you can do now
If you are a resident, fellow, or early-career proceduralist
Start your own copy now, in parallel with whatever institutional system you are required to use. Do not assume you will have access to your case log after you leave your current program or job; the Cadish 2016 data alone shows the institutional version is already unreliable, and access policy after departure is inconsistently published across specialties.
Ask your supervising attendings for something more than a system-generated case count. A brief, specific note, even outside any formal system, describing what you did and that they watched you do it, is closer to a real attestation than a CPT code will ever be, and it costs almost nothing for a generous attending to provide.
Request an export of your ACGME case log data before you graduate, in whatever format is available, rather than assuming continued access afterward.
If you are a mid-career or senior physician
Start reconstructing your own record now, before the next job change forces you to do it under time pressure. The 5-to-20-hour reconstruction estimate assumes it happens once, under deadline. Building your own running record incrementally, over years, is dramatically cheaper than doing it all at once during a privileging crunch.
Offer to countersign for former trainees you supervised. You are exactly the kind of attestation source this system currently has no place for, and a specific, dated note confirming what you personally observed is a real gift to someone reconstructing their own record years later.
If you lead a medical staff office or credentialing committee
Ask what your institution's actual policy is on data access for departing physicians, and whether it is written down anywhere accessible. The underlying research for this article found that policy in this area is inconsistent and rarely published, which means your own physicians likely do not know what to expect when they leave.
Consider what evidence format you would actually accept from a peer-countersigned external log. The value of any future portable-record system depends entirely on institutional willingness to treat it as legitimate supporting evidence, and credentialing committees willing to say so publicly are a precondition for that system existing at all.
If you lead a specialty society or accrediting body
Study the UK's eLogbook model directly. The UK's Intercollegiate Surgical Curriculum Programme built a surgeon-owned, portable electronic logbook that travels across NHS trusts, functioning as the closest working counterexample to the fragmented US approach, and its usage data is directly relevant to whether US physicians would sustain voluntary logging given genuine ownership of the record.
Frequently asked questions
Can I access my ACGME case logs after residency? Policy on post-graduation access to the ACGME Case Log System varies by specialty and is not published in a single, accessible, comparable form across specialties, according to the underlying research for this article. Physicians should request an explicit export of their data before graduating rather than assuming continued access afterward.
How do new attendings prove case volumes for hospital privileges? Typically by requesting historical records from former employers' EHR systems, personal spreadsheets, and specialty board case lists, a fragmented process that the underlying research estimates costs 5 to 20 hours per job change for the roughly 90,000 residents, fellows, and job-changing physicians affected annually.
How accurate are resident case logs? Not very, by residents' own accounts and by objective measurement. Only 11.3 percent of surgical residents rate their case logs as highly accurate, and general surgery residents' actual coding accuracy measures 52.82 percent even though 74 percent believe they are logging accurately, according to Cadish et al. (2016) and Balla et al. (2016), both published in the Journal of Surgical Education.
What counts as evidence of procedural competence for OPPE or FPPE? Currently, primarily institution-generated volume data from EHR reports and, where available, specialty board case lists, since no standardized, portable, peer-attested logbook format exists across US institutions. Ongoing professional practice evaluation and focused professional practice evaluation processes depend on volume history that is difficult for physicians to independently reproduce.
Who owns a physician's teaching record? In practice, the institution's GME office, through systems such as MedHub or New Innovations, which are not built to export a portable record when the clinician-educator changes institutions. No equivalent to a case log exists for teaching and mentorship contribution, despite promotion and academic advancement depending heavily on exactly this kind of longitudinal record.
Why don't hospitals just share case volume data when a physician changes jobs? There is no established data-sharing standard or shared incentive between hospitals for this purpose; each institution's EHR and credentialing systems are built to serve that institution's own accreditation and liability needs, not a departing physician's next employer, and hospitals generally have little structural incentive to invest resources helping a departing physician assemble a portable record.
The bottom line
The surgeon spending two weeks emailing former administrators is not dealing with a bureaucratic inconvenience. She is dealing with the direct consequence of a system that never decided a physician's own procedural history belongs to the physician, the way a century-old aviation norm decided a pilot's flight hours belong to the pilot.
The record was never solid to begin with. Only 11.3 percent of residents trust their own case logs, and the actual coding accuracy behind those logs runs as low as 58 percent in some specialties, badly miscalibrated against the near-universal belief that the numbers are right. Then, at graduation and at every job change after that, the fragile record gets orphaned entirely, split across institutions that have no reason to help reassemble it and no format for the one thing that actually matters: a named, verified peer's attestation that they watched this specific person do this specific work.
Forty thousand residents graduate a year. Fifty thousand physicians change jobs. Each of them pays this tax again, in hours and in delayed credentialing, because ACGME's remit ends at graduation, boards keep records only long enough for exam eligibility, hospitals have no reason to help a departing employee, and no medical society has built a career-spanning, individually owned, peer-signed alternative.
A pilot's logbook has followed her through every job change of her career, signed, continuous, and entirely hers. A surgeon fifteen years into practice is, right now, cold-emailing former colleagues, trying to prove she did the job she has already spent a decade and a half doing.
Part of a series on the missing professional infrastructure of healthcare. Previously: Navigational Capital
Evidence note: case-log accuracy and burden figures are from Cadish et al., Journal of Surgical Education, 2016 (PMID 27049679); coding accuracy figures are from Balla et al., Journal of Surgical Education, 2016 (PMID 27886974), and Collins et al., Journal of Surgical Education, 2020 (PMID 32950429). The 112-day figure for contract-to-start-date delay is drawn from AAPPR data as reported by the American Medical Association; incomplete credentialing verification is a commonly cited contributor to that gap rather than a formally isolated causal factor in the cited source. The $70-280 million annual cost-of-reconstruction estimate is a calculation built from published population figures (roughly 40,000 graduates and 50,000 job-changing physicians annually) and an assumed 5-to-20-hour reconstruction time and $150 hourly opportunity cost; the hours and dollar figures are reasoned estimates from the underlying research dossier, not a directly measured survey result, and should be read accordingly. The claim that ACGME graduate-access policy "varies by specialty and is not published in accessible form" reflects the underlying research's own characterization of an information gap, not a specialty-by-specialty audit conducted for this article. Comparisons to the UK's Intercollegiate Surgical Curriculum Programme eLogbook are offered as a structural counterexample and have not been independently verified against current NHS documentation in this pass.
