THE INTERPROFESSIONAL
By Peter Schindler, MD, PhD
← All Essays
Issue 038·November 18, 2026·Clinical Practice

Handoff as Epistemology: What Cannot Be Said at Shift Change

Handoff moves what a clinician can say. The most valuable thing she knows is frequently the part she cannot say, and the form of the handoff selects against it.

10 min read · By Peter Schindler, MD, PhD

Nothing is handed off at handoff.

What crosses is a set of sentences. The outgoing clinician's knowledge stays with the outgoing clinician, walks to the parking garage, and drives home. What the incoming clinician receives is a description of that knowledge, produced under time pressure, shaped by what the form will accept, and stripped of everything the sender could not put into words. She does not inherit knowing. She inherits a report about knowing, and she has to build the knowing again herself, from the beginning, in the same room.

Issue 016 of this publication argued that sign-out is a genre: a structured communicative form with conventions, recognized failure modes, and a research literature, and that protocols like I-PASS and SBAR improve it. That argument holds. This one sits underneath it. The genre question is what shape the sentences should take. The epistemological question is which kinds of clinical knowledge can be put into sentences at all, and what becomes of the kinds that cannot.

Gilbert Ryle drew the line in The Concept of Mind in 1949. Knowing-that and knowing-how are different in kind. Knowing that the potassium is 3.1 is one sort of thing. Knowing how to tell, from the doorway, that a patient's breathing has changed in a way no number has yet registered is another. Ryle's argument was that the second is not a compressed version of the first. It does not decompose into propositions without remainder.

Handoff is a machine built almost entirely for knowing-that. It moves numbers, events, plans, and contingencies, and it moves them well. The overnight course, the pending culture, the threshold at which to call: these travel intact, and structure makes them travel better. That is not a small achievement. Issue 016 reported the reductions in medical error and preventable harm that the 2014 I-PASS trial produced. Structure works, and what it works on is the transmissible part.

The problem is the remainder.

Issue 007 of this publication argued that much of what an experienced nurse knows is tacit in Michael Polanyi's sense, held in the body and barely articulable by the nurse who holds it. In clinical work that gap between knowing and telling has a shift change in it. The nurse at 0645 knows more about the patient in 412 than she can say in the four minutes she has to say it. Some of what she cannot say is trivial. Some of it is the most valuable thing she holds.

The Filter

Handoff does not merely fail to carry tacit knowledge. It selects against it.

A handoff is a public performance of clinical reasoning. Someone is listening, and that someone can ask why. The genre rewards claims that can be supported: here is the finding, here is the trend, here is why I am worried. It penalizes claims that cannot. Issue 004 of this publication argued that the nurse's running model of a patient is a real clinical signal, and that the flowsheet carries her data without carrying her model. A nurse who says "I do not like how she looks" is saying the model out loud. Asked what the vitals are, and able to answer only that the vitals are fine and she still does not like how she looks, she has by the standards of the form said nothing at all. She has failed to produce evidence.

She has not failed to know something. Issue 017 of this publication set out Benner's five stages and the character of the last one: the expert's perception arrives whole, and asked to explain it she frequently cannot. Handoff demands the explanation anyway, and what it gets is a rationale that was not the route she actually took. The rationale is a translation produced afterward, and frequently a poor one.

So the sentence gets dropped, or downgraded into something the form will take. "I do not like how she looks" becomes "she had a restless night." Restless night is receivable. It fits. It is also nearly empty: it names an event and discards the perception that made the event worth mentioning.

The filter runs in one direction. It removes the highest-value, least-justifiable content and preserves the lowest-value, most-justifiable content. It does this at the precise moment a patient's care changes hands, which is the moment the perception was most needed.

The same filter runs inside medicine, and it runs on physicians handing off to physicians. An attending who has followed a patient for three days holds a sense of trajectory that never reaches her note, because the note is built to hold findings and plans. She signs out on Friday and says the patient is improving. What she does not say, because there is no sanctioned place to put it, is that this improvement has the wrong texture, that it resembles the improvements she has watched reverse. Her colleague inherits improving. The reservation stays behind with the person who had it.

No protocol corrects this, because the filter is the demand for justification, and a protocol is how that demand gets made explicit. A field labeled situation awareness invites the clinician to name what she is worried about and why. She can fill it. She cannot fill it with something she is unable to articulate, and a field left thin because there was nothing sayable looks identical to a field left thin because there was nothing there.

The Asymmetry

The outgoing clinician knows the shape of what she left out. The incoming clinician does not.

This asymmetry is the specific epistemic hazard of handoff, and it grows rather than shrinks as handoff becomes more structured. The sender is aware of her own compression. She knows she skipped the daughter's phone call, the fifteen minutes the patient spent staring at the ceiling, the three extra times she checked on 412 for a reason she could not name. She experiences the handoff as a summary, because she still has the original.

The receiver has no original. He has the summary and nothing to measure it against. Omissions have no shape from the receiving end. There is no perceptual difference, for him, between a fact never mentioned because it did not exist and a fact never mentioned because it could not be said.

A completed structured handoff makes this worse by looking finished. Every field is populated. Illness severity assigned, action items listed, synthesis by receiver performed and confirmed. The document is formally complete, and formal completeness reads as epistemic completeness. The receiving clinician's confidence is now calibrated to the form rather than to the patient.

That is not an argument against structure. It is an argument that structure changes the failure mode instead of removing it. The unstructured handoff fails as nobody told me. The structured handoff fails as I thought I had been told everything. The second failure is quieter and takes longer to find.

The scene that follows is fictional.

A woman of seventy-eight is two days out from surgical repair of a hip fracture. Her vital signs across the night are unremarkable. She is oriented when asked. Her night nurse has been into the room more often than the assignment required and, if pressed, could not say why, other than that the patient answered questions a beat slower than the night before and had stopped asking about going home.

At 0700 the nurse hands off. She has eleven minutes for four patients. What she says about this one is: post-op day two, hip, vitals stable, oriented, slept poorly. Every word is true. The handoff is accurate and complete by the standards of the form.

By early afternoon the patient is pulling at her lines and does not know the year. The signal existed at 0400. It existed in a nurse who kept returning to a room for a reason she could not state. It did not survive the sentence.

What Can Be Made Transmissible

That scene is not an argument that tacit knowledge is permanently untransferable. It is an argument that transferability has to be built, and that building it is conceptual work rather than communication work.

Sharon Inouye's Confusion Assessment Method, published in the Annals of Internal Medicine in 1990, is what success looks like. Before the CAM, the perception that an older patient was not herself was a private impression with no institutional standing. Inouye and her colleagues did the work of decomposing that impression into four features: acute onset with fluctuating course, inattention, disorganized thinking, and altered level of consciousness. The instrument did not create the knowledge. Bedside nurses had it already, and had it long before anyone asked them for it. What the instrument created was a public form the knowledge could take. A way for the perception to survive a shift change, enter a note, and be argued with.

The instrument was designed for clinicians without psychiatric training, which is the point. It was built so the observation could move between people who had been taught differently.

Even so, the quiet form of delirium is still missed. The patient who is withdrawn rather than agitated does not trigger the category, because withdrawn looks like tired and tired looks like a person recovering from surgery. That is a fact about the vocabulary, not about the attentiveness of the person at the bedside. The perception is available. The name for it arrives late or not at all.

This is the usable lesson. When a category of clinical knowledge repeatedly fails to survive handoff, the answer is rarely a better protocol and never a longer one. The answer is to give the perception a name both professions recognize, so that saying it costs nothing and hearing it means something.

Issue 012 of this publication set out Risjord's thesis that nursing's position in clinical structure generates knowledge that is not derivative of medical knowledge. That thesis matters here in a way that is easy to miss. If the two positions generate different knowledge, then handoff between a nurse and a physician is not a transfer within one epistemic system. It is a translation across two. The vocabulary that would make the nurse's perception fully legible to the physician frequently does not exist, because the two professions built their vocabularies separately and without reference to each other. The CAM is one of the rare cases where such a vocabulary was constructed on purpose.

Christine Tanner's clinical judgment model, published in the Journal of Nursing Education in 2006, begins with noticing, and makes noticing conditional. What a clinician notices depends on her background, on the context, and on her relationship with this particular patient. Tanner, with Benner, Chesla, and Gordon, had earlier named this knowing the patient, and distinguished it from knowing about the patient's disease.

The incoming clinician does not have it. At the moment the handoff ends he knows about the patient and does not know the patient. He therefore cannot yet notice. The perceptual apparatus for this person has not been built, and it does not arrive in the report.

That settles the posture a handoff deserves. It is a hypothesis, not a record. It is the best available account of a patient, produced by someone who has now left the building, and its function is to tell the receiving clinician where to look first rather than what is true.

Nursing has an institutional version of this. Bedside shift report, which moves handoff into the patient's room with the patient present, is recommended in the Agency for Healthcare Research and Quality's Guide to Patient and Family Engagement in Hospital Quality and Safety, and many units have adopted it. The usual justification is patient engagement. The epistemic justification is stronger. At the bedside both clinicians can point. The outgoing nurse can say look at her and the incoming nurse can look. The sentence is checked against its referent in real time. Away from the bed, the sentence is the only referent there is.

Issue 024 of this publication argued that the move from bedside rounds to hallway and table rounds gained efficiency and lost something pedagogical and relational. The handoff version of that trade is epistemic. Physician sign-out almost never happens in the room. It happens in a workroom, in front of a list, about people none of whom are present, and every claim made about them is made about an absent object.

Two obligations follow, and both are uncomfortable.

The sender's obligation is to say the unjustified thing and to label it as unjustified. I have nothing to point at and I am worried about her is a complete and useful handoff sentence. It reports the state of a trained perceptual system, which is clinical data. It requires a unit culture in which an unsupported worry is receivable content rather than evidence that the speaker cannot reason. Issue 003 of this publication introduced Donald Schön through reflection-in-action and the reflective journaling nursing curricula built on it. The claim underneath that concept is the one that bears here. The Reflective Practitioner (1983) argues that professions organized around technical rationality have nowhere to put practitioner knowledge that will not resolve into technique. Handoff is where having nowhere to put it becomes expensive.

The receiver's obligation is to go and look early, before the report hardens into belief. Not to verify the numbers. To acquire first-hand the thing that could not be sent.

The handoff ends. The outgoing clinician goes home carrying whatever she knew and could not say. In the room there is a patient, and a clinician who has just been told about her. Nothing was handed over. A description changed hands. The knowing has to be built again, by someone else, starting now.

About the author

Peter Schindler, MD, PhD is an Assistant Professor of Medicine and Associate Program Director of the Community Health Center Family Medicine Residency Program at the University of Nebraska Medical Center. He practices at Winnebago Comprehensive Health System, OneWorld Community Health Center, and Nebraska Medicine. He completed a Primary Care Research Fellowship at McGill University and holds a BSN from the University of Wisconsin-Oshkosh, an MS and PhD in nursing from Emory University's Laney Graduate School, a Diploma in Tropical Medicine from the Liverpool School of Tropical Medicine, and an MD from the Medical College of Wisconsin-Green Bay. He is board certified by the American Board of Family Medicine. The Interprofessional publishes new essays every week at the intersection of medicine, nursing, and the clinical knowledge that lives between them.

Disclaimer. The views in The Interprofessional are Peter Schindler's own and do not represent the official positions of the University of Nebraska Medical Center, Winnebago Comprehensive Health System, OneWorld Community Health Center, Nebraska Medicine, or any other affiliated institution.