A cardiologist sits down on a Friday evening and, out of curiosity more than habit, tries to count her week.
Four curbside questions from former trainees, texted or caught in a hallway. Two hours of informal peer review for a colleague's tricky case. Three fellows she has been quietly mentoring, unofficially, because none of them were formally assigned to her. A resident's difficult conversation with a family that she walked through, step by step, the night before it happened. A "what would you do here" message from a colleague at another hospital entirely, answered in four sentences between patients.
None of it will appear anywhere that counts. Not in her CV. Not in her RVU report. Not in the promotion packet she is quietly dreading assembling next year, which will ask her to list publications and grants and will offer, at most, a thin "teaching statement" section that reviewers routinely discount against a PubMed-verifiable line.
She is not confused about whether this work matters. She knows, the way every clinician who has been doing this long enough knows, that it is the actual mechanism by which judgment gets transmitted in medicine: not the textbook, the curbside. Not the CME module, the fellow standing next to her during a hard case.
The formal record of her career will capture what she published and billed, and will contain no trace of the work she is most confident actually made other clinicians better, because nothing in medicine's record-keeping was ever built to see it.
The scale of the thing nobody counts
Start with how common this behavior actually is, because it turns out to be close to universal rather than a niche academic habit.
A 2021 study published in BMC Family Practice found that 95 percent of surveyed family physicians report requesting informal medical advice from colleagues. Eighty-one percent cited availability and accessibility as the reason they reach for an informal consult rather than a formal one.
Here is the finding that reframes the whole picture: only 11 percent felt satisfied giving informal consults to family or friends, and 49 percent reported discomfort doing so. Sixty-six percent said they wanted formal guidance on the practice.
Read those together. This is not a small, edge-case behavior a handful of physicians engage in occasionally. It is something close to universal, done constantly, by people who are simultaneously uneasy about it and receiving no institutional support for how to do it well. That combination, near-universal practice plus real discomfort plus explicit demand for guidance, is the signature of an activity that has outgrown the infrastructure built around it.
Why the formal record looks the way it does
Assessment for hiring, promotion, and tenure across academic medicine is dominated by publication and grant metrics, a pattern documented plainly by Moher and colleagues in PLoS Biology in 2018. That dominance persists despite wide acknowledgment, including from the researchers studying it, that it crowds out teaching, mentorship, and service contributions.
This is not a mystery once you look at why publication and grant metrics won that competition in the first place. They are legible. A citation count is a number. A grant is a dollar figure with a start and end date. Teaching, mentorship, and curbside judgment produce none of that. They leave no artifact that a bibliometric database, a GME office spreadsheet, or a promotion committee's standard rubric was ever built to capture.
The formal record measures what was built to be legible, not what was valuable. Nobody designed medicine's record-keeping infrastructure to deliberately exclude teaching and mentorship. It simply was never asked to include something that, by its nature, happens in a hallway, leaves no chart note, and exists only in the memory of the two people involved.
What the current workflow actually captures
Walk through where a physician's contribution supposedly gets recorded today, and notice how thin each channel actually is.
Teaching hours get self-reported into a GME office spreadsheet or a system like MedHub, with no independent attestation from the learners who were actually taught. Mentorship becomes visible, if it becomes visible at all, only if a mentee happens to write a thank-you line in a paper's acknowledgments section years later. Curbside consults leave no trace by design, which is precisely why physicians use them: fast, informal, no chart note, no bureaucratic overhead.
Annual review and promotion-and-tenure dossiers lean on a self-assembled "teaching portfolio," a document the faculty member builds themselves, without independent verification, that reviewers routinely weight less heavily than a line on a CV that a database can independently confirm.
Every one of these channels shares the same structural weakness: it depends on self-report, after the fact, with no counterparty who can attest that the contribution actually happened the way it is described.
Why nobody owns it
Look at who could plausibly fix this, and why each candidate's business or mission stops short.
Bibliometric databases like Scopus and Web of Science monetize publication counting. Counting citations is their product. Building a parallel system to count and verify teaching contributions is not adjacent to that business; it is a different business entirely, with no revenue model attached.
GME offices track teaching hours for accreditation purposes, which is a compliance function with a specific, narrow scope. Accreditation-relevant hour-tracking was never designed to produce a portable credit system a physician could carry from one institution to the next.
Doximity, despite building substantial physician-network infrastructure, has no attestation product connecting two named individuals around a specific act of mentorship or teaching. Its revenue model runs through pharma-facing attention, not peer-to-peer verification.
Professional societies give teaching awards, typically to a small handful of people a year, based on nomination. A nomination-based award recognizes visibility and self-promotion as much as actual volume of contribution, and by design it can only ever recognize a tiny fraction of the people doing the work.
No commercial actor's revenue depends on making informal contribution visible. That is the whole explanation for why, despite near-universal recognition that this work matters, nobody has built the infrastructure to see it.
The hidden graph
Here is a way to think about what is actually missing, structurally.
Every curbside consult, every teaching session, every mentorship relationship is, in a real sense, an edge in a graph: taught-by, mentored-by, consulted-by, connecting two specific people around a specific act of professional contribution. That graph exists. It is enormous. It is also entirely unwritten, living only in the memories of the two people involved in each edge.
No institution, including the ones with a genuine incentive to reward this behavior (a department chair who wants to retain a valuable but underpaid clinician-educator, a promotion committee that says it values teaching), can currently query that graph. They cannot ask "who is actually teaching and mentoring the most in this department," because the edges were never recorded with a counterparty who could independently attest to them. The information a chair would need to make that judgment does not exist as data anywhere, even though it exists, richly, as lived experience.
Why this is a data structure problem, not a values problem
It is tempting to read this as an institutional failure of values: medicine says it values teaching and does not actually reward it. That framing is not wrong, but it understates how solvable the problem actually is.
Medicine already runs on an informal contribution economy larger, in raw volume, than its formal one. The 95 percent figure above is not a niche behavior; it is closer to a professional universal. The problem is not that this work lacks value in anyone's eyes. It is that it lacks a data structure: a way to record who did what, for whom, with a counterparty who can verify it happened.
A peer-attested contribution ledger does not need to change anyone's incentives. It needs to make an already-valued behavior visible to the people, department chairs, promotion committees, mentors themselves, who would act differently if they could actually see it.
What would actually work
Countersigned entries, not self-report. A contribution record only carries weight if the counterparty, the trainee who was taught, the colleague who received the curbside answer, can confirm it happened. Self-reported teaching hours are exactly the weak signal that promotion committees already discount; a countersigned entry is a different kind of evidence entirely.
Narrative-light by design. The record should capture that a contribution happened, its type, and its counterparty, not a detailed account of what was discussed. This keeps the ledger fast enough that people actually use it and avoids creating a new documentation burden that defeats its own purpose.
Never a vehicle for clinical content. A curbside consult about an actual patient must never be logged with patient identifiers or clinical detail. The record is about the relationship and the contribution type, full stop; anything more creates liability exposure that would kill adoption instantly.
Opt-in, not surveillance. Both the contributor and the counterparty need to consent to logging a given interaction. A system that silently tracks every hallway conversation would collapse the exact informal, low-friction quality that makes curbside consultation valuable in the first place.
Exportable in a format promotion committees actually use. The ledger's value is realized only when it produces something usable: an annual summary formatted for a P&T dossier, not a raw data dump a reviewer has to interpret themselves.
Verified membership on both ends. A countersignature only means something if both parties are confirmed to be who they say they are. Anonymous or unverifiable attestation reintroduces the exact self-report weakness the system is meant to fix.
Built for cognitive and procedural specialties differently. In procedural specialties, a substantial share of teaching happens through supervised cases, which have their own existing logbook conventions. In cognitive specialties, teaching is almost entirely curbside and currently undocumented in any form. A single ledger design needs to accommodate both without forcing one specialty's format onto the other.
What you can do now
If you are a clinician-educator
Start your own informal log today, even without any platform to support it. A running list of who you taught, mentored, or consulted with informally, and when, is worth having in your own files regardless of whether any institutional system ever catches up.
Ask mentees and colleagues directly for a written acknowledgment. A short email confirming "thanks for walking me through X" is a countersignature in its simplest possible form, and it costs the other person almost nothing to send.
Bring the data, not just the feeling, to your annual review. If you can quantify even roughly how many informal consults or mentoring conversations you handled in a year, that number is more persuasive to a committee than a general statement that you "value teaching."
If you sit on a promotion and tenure committee
Ask explicitly, in every dossier review, what informal teaching and mentorship a candidate does that the packet does not show. The near-universal prevalence of curbside consultation (95 percent, per the cited survey) means almost every candidate in front of you is doing more of this work than their dossier reflects.
Weight a countersigned account more heavily than a self-reported one, even an informal one, since a confirmed contribution is meaningfully stronger evidence than an unverified claim, regardless of how polished either document looks.
If you build systems
Design the attestation flow before the analytics. The core defensible asset here is the verified, countersigned record itself; dashboards and benchmarking reports are valuable downstream products, but they depend entirely on getting the underlying attestation mechanism right, and low-friction, first.
Keep the schema deliberately narrow. A contribution-type field, a date, a counterparty, and an optional short narrative is enough. Anything closer to full clinical or performance documentation invites both a liability problem and a usage problem, since people will not log something that feels like being monitored.
Frequently asked questions
How do academic physicians get credit for teaching in promotion decisions? Mostly through self-assembled teaching portfolios and self-reported hours logged in systems like MedHub, which promotion committees routinely weight less heavily than publication metrics that can be independently verified through a database like PubMed or Scopus. Assessment for academic hiring and promotion is dominated by publication and grant metrics, according to Moher et al., PLoS Biology, 2018.
Is it common for physicians to give informal medical advice to colleagues? Yes, close to universal. A 2021 study in BMC Family Practice found 95 percent of surveyed family physicians report requesting informal medical advice from colleagues, with 81 percent citing availability and accessibility as the reason (Amran et al., BMC Family Practice, 2021).
Why are clinician-educators sometimes promoted less often than research faculty? Because promotion and tenure evaluation systems are built around legible, verifiable metrics like publications and grants, and teaching, mentorship, and informal consultation produce no equivalent artifact. This is a documented pattern in the literature on academic medicine assessment (Moher et al., PLoS Biology, 2018), though this article did not locate an independently verified, primary-sourced statistic directly comparing promotion rates between clinician-educator and research tracks, and does not claim one.
What counts as service in a physician promotion dossier? Service typically includes committee work, mentorship, teaching, and informal consultation, but unlike publications and grants, these contributions usually rely on self-report rather than independent verification, which is a significant reason they carry less weight in most current promotion evaluations.
Are physicians comfortable giving informal curbside consults? Not entirely. The same 2021 BMC Family Practice survey found only 11 percent of physicians felt satisfied giving informal consults to family or friends, 49 percent reported discomfort doing so, and 66 percent said they wanted formal guidance on the practice, despite 95 percent doing it anyway.
How could a contribution ledger work without creating patient privacy or liability problems? By logging only the relationship and contribution type, such as "mentored a fellow through fellowship applications" or "answered a colleague's clinical question," with a countersignature from the other party, and never logging patient identifiers or clinical case detail. This keeps the record focused on professional contribution rather than becoming an informal, undocumented clinical record.
The bottom line
The cardiologist counting her week on a Friday evening is not wrong about what mattered most in it. The four curbside consults, the informal mentorship, the walk-through before a hard family conversation, these are not a distraction from her real job. For a huge share of practicing clinicians, per the 95 percent figure cited above, this near-invisible work is a constant, close to universal part of how medicine actually functions day to day.
None of it will appear in her promotion packet next year, not because anyone decided it does not matter, but because nobody ever built a way to record it that a committee could trust. Publications and grants won the competition for institutional attention because they produce numbers a database can verify. Teaching, mentorship, and curbside judgment produce nothing a database was ever asked to see.
That is a solvable kind of problem, and a more tractable one than it first appears. Medicine does not need to convince anyone that this work has value; the near-universal prevalence and the explicit demand for guidance (66 percent, in the cited survey) already establish that. What it needs is a data structure: a countersigned, narrative-light, opt-in record that turns an already-valued behavior into something a chair or a committee can actually query.
Until that exists, the graph of who taught whom, who mentored whom, who answered whose hard question at nine at night, stays exactly where it has always lived: in the memory of the two people who were there, invisible to everyone else, and gone the moment either of them forgets or moves on.
Part of a series on the missing professional infrastructure of healthcare. Previously: The Second Opinion 5 Billion People Can't Get
Evidence note: the informal-consultation prevalence and satisfaction figures (95 percent request informal advice, 81 percent cite availability, 11 percent feel satisfied giving informal consults to family or friends, 49 percent report discomfort, 66 percent want formal guidance) are from Amran MM, Kopit AB, Kranc HA, Peleg R, BMC Family Practice, 2021. The finding that academic assessment is dominated by publication and grant metrics, crowding out teaching and service, is from Moher D et al., PLoS Biology, 2018. This article's dossier source flagged the specific claim of a clinician-educator versus investigator-track promotion-rate disparity as "unverified, exact primary source not independently located," and this article presents that point accordingly, as a documented pattern in the broader literature rather than a single confirmed statistic. Nothing in this article is guidance on documenting any specific patient interaction; any contribution-ledger concept described here explicitly excludes clinical content and patient identifiers.
