---
title: Read in a crisis, misread in court
author: Lee Farmer
url: https://www.plaintiffmagazine.com/issues/2026-october/read-in-a-crisis-misread-in-court
---

# [Read in a crisis, misread in court](https://www.plaintiffmagazine.com/issues/2026-october/read-in-a-crisis-misread-in-court "Read in a crisis, misread in court")

## What emergency anesthesia manuals do and don’t tell us about the standard of care

[Andrea J. Traynor](https://www.plaintiffmagazine.com/authors/andrea-j-traynor "Andrea J. Traynor")

[Larry F. Chu](https://www.plaintiffmagazine.com/authors/larry-f-chu "Larry F. Chu")

[2026 October](https://www.plaintiffmagazine.com/issues/2026-october)

**The 2:14 a.m. problem**

The clock on the anesthesia workstation reads 2:14 a.m. The case is an exploratory laparotomy, an urgent open-abdominal operation for a bowel obstruction; the patient is 61, and until this moment everything has been routine. Then the pulse oximeter’s tone begins to fall, its pitch dropping as the oxygen level in the patient’s blood drops. Saturation 94, then 88, then 81. Airway pressures climb; it is getting harder to push air into the lungs. The blood pressure cuff cycles and comes back at 62 over 38. The surgeon asks a question nobody answers. Over the next four minutes, while the numbers keep moving, the anesthesiologist will make 11 consequential decisions: hand-ventilate or swap the breathing circuit, epinephrine or vasopressin, fluid or blood, call for help or manage alone. On the wall beside the workstation, in a laminated binder nobody has touched, hangs an emergency manual.

Eighteen months later the case has a caption, and the first exhibit both sides reach for is that binder. The plaintiff’s argument: the team skipped steps the manual required, and deviation is negligence. The defense’s argument: the team followed the manual, so the care met the standard. Both arguments rest on the same misunderstanding of what the document is.

A cognitive aid, the clinical term for the checklists and emergency manuals designed to be read during a crisis rather than memorized before one, was built to support a clinician’s memory in real time. It was not built to describe what reasonable care requires, and reading it as though it were produces conclusions the evidence will not support in either direction.

We write as authors and/or editors of these manuals, and we write as physicians, not lawyers; nothing in this article is legal analysis. What follows is what we know about these documents from the inside: where they come from, how they change, what they are engineered to do, and what a record actually shows about whether one was used. Those facts raise legal questions, about admissibility, about the weight of internal policies, about causation, and those questions belong to counsel. The facts themselves are ours, and we can defend them under oath.

When used correctly, an emergency manual can raise more significant issues than whether a doctor merely “checked every box.” It may also show what safety systems existed, how the hospital planned for foreseeable human error in medical emergencies, whether the staff had been trained to use the manual, and whether the system worked as it was supposed to when the patient needed it most.

**Why lawyers love these documents**

The appeal is real, and it is bipartisan. In a specialty where the standard of care must ordinarily be built from dueling expert opinion, a cognitive aid looks like the standard of care written down: a document that arrives pre-credentialed by an institutional name and an aviation pedigree, and that a lay jury can hold, read, and apply. It converts a medical question into a matching exercise. Nobody in the jury box can say whether vasopressin was reasonable at minute three, but anyone can compare a numbered list against a chart. And it is a cross-examination machine, because printed steps generate the closed, controlled questions trial lawyers prize. (“Step 4 says call for help. You didn’t call for help, did you?”)

Each side then adds its own reasons. For plaintiffs, deviation makes a simple, devastating story: the answer was hanging on the wall, and nobody opened it. That story converts a judgment case into a rule-breaking case, and it comes with a built-in aviation analogy for jurors who have all been passengers. For the defense, adherence feels like a safe harbor, and a wall of manuals plus a calendar of simulation training reads as a safety-conscious institution whose bad outcome was not bad care.

Underneath all of it sits one promise: that the manual can convert the hardest question in a crisis case, what a reasonably careful anesthesiologist would have done in those four minutes, into the easiest, whether the chart matches the page. That promise is not one these documents can keep. What they can do is different, and in our view more useful: they tell you a great deal about the system the clinician was working inside.

**What a cognitive aid is designed to do**

Emergency manuals descend from aviation. When an engine fails, a flight crew does not brainstorm; it runs a checklist, because decades of accident investigation established that even excellent pilots omit critical steps under stress. Anesthesiology, a specialty that has long modeled its safety culture on aviation’s, borrowed the insight and built crisis checklists for the operating room, among them the Stanford Emergency Manual, a set of algorithms for operating room emergencies from anaphylaxis to malignant hyperthermia (a rare, rapidly lethal reaction to anesthetic drugs) that now hangs in operating rooms around the world.

Many readers will know this movement from Atul Gawande’s 2009 bestseller, The Checklist Manifesto. Four of its premises matter for anyone reading one of these documents in a record: complex systems fail because human memory and attention are limited; even experts benefit from checklists because expertise does not eliminate predictable failures of memory and attention under stress; good checklists are short, focused, and used at critical points; and a good checklist supports teamwork and judgment rather than replacing them. Anesthesiology was among the first fields to embrace these concepts. David M. Gaba pioneered simulation-based crisis resource management, William Berry helped lead the development of the World Health Organization’s Surgical Safety Checklist, and the Stanford Anesthesia Cognitive Aid Group created emergency manuals for the operating room.

The field also distinguishes how a checklist is meant to be used. In read-do, a designated reader calls out each step as the team performs it. In do-confirm, work performed from memory is verified afterward. Crisis manuals like the Stanford Emergency Manual are built for read-do use, in real time, under stress. That distinction matters more than any individual step on any individual page, because it tells you what using the manual correctly actually looks like, and therefore what its use or non-use should have left behind in the record.

The design premise is specific: crises cause retrieval failure, not knowledge failure. The anesthesiologist who cannot summon the next step of the malignant hyperthermia protocol at 2 a.m., with alarms sounding and a surgeon calling out, knew it cold at her board examination. The manual exists for the 2 a.m. brain, and its design says so: large type readable across a room, one crisis per page, steps ordered by urgency rather than textbook logic, and an explicit assumption that someone other than the physician managing the crisis will read the steps aloud. A document engineered to jog a saturated memory in ninety 90 was not engineered to define a professional standard.

The simulation literature establishes the same premise from the other direction. Put excellent, well-trained teams into a high-fidelity simulated crisis and predictable things happen: fixation error, in which the team persists with a failing plan; loss of situational awareness; omission of steps every member can recite in the debrief an hour later. In a randomized simulation trial from Gawande’s research group, published in the New England Journal of Medicine, operating-room teams managing simulated crises failed to perform 23% of lifesaving process steps when working from memory, versus six percent when a checklist was available. Every team performed better with the checklist in hand. (Arriaga et al., Simulation-Based Trial of Surgical-Crisis Checklists (2013) 368 New Eng. J. Med. 246.)

We want to be precise about what that finding does and does not say, because it is regularly read too fast in both directions.

It does not say the manual defines correct care. The checklist arm still missed six percent of steps, and the study measured process compliance in simulation, not patient outcome in an operating room.

It does say that skilled, credentialed, well-trained clinicians predictably omit a substantial fraction of critical steps when they work from memory alone, and that the omission rate falls sharply when a checklist is available and used. This is not a marginal or contested finding in our field. It is the reason these manuals exist, and every institution that has adopted one has, in adopting it, accepted the premise. Whatever legal significance that carries is for counsel to assess. In clinical terms, a crisis record that reveals omitted steps, on its own, does not prove that a clinician is atypical. Rather, it fits the foreseeable cognitive lapses documented in skilled clinicians operating under crisis conditions, shifting attention to what the institution did to support the predictable failure it had already recognized.

**What a functioning crisis system looks like**

Crisis resource management is a trained discipline with specified structure, and this is the part of the picture most often missing from a record.

A functioning system assigns roles before they are needed. An event manager keeps the whole room in view rather than performing tasks. A designated reader owns the manual and nothing else, calling out steps and confirming execution. Help is summoned early, by a defined mechanism, rather than by someone eventually noticing that more hands are needed. The manual is physically present where the anesthetic is being given, in the edition currently in force. The department has trained on the scenario, and can say when, and who was in the room.

When that system is working, it leaves traces. A reader is identifiable in testimony. Steps appear called and confirmed. Paging and code-activation logs show help requested at a time consistent with the physiology. Roles are described consistently by people who were not standing next to each other.

When it is not working, the absence is also visible. No one can say who was reading. No one can say where the binder was. The training program exists as a policy without an attendance roster. The edition on the wall is two revisions behind the one produced in discovery.

Neither pattern is self-interpreting, and we would caution against treating the second as a conclusion. Departments with excellent systems have bad nights, and role structure can operate informally among people who have worked together for years without generating a paper trail. But the distinction between an individual clinician’s performance and the system built around that clinician is a real one in our field, it is the distinction the safety literature is largely about, and a case evaluated without it is missing most of the picture. How institutional adoption, training programs, and crisis resource management implementation bear on institutional versus individual responsibility is a legal question we are not qualified to answer. What we can say is that the record usually contains the raw material for asking it, and that most requests never gather it.

**Version and lineage**

Emergency manuals are often treated as fixed objects. They are not. They are living documents with a publication history, and the history matters.

Consider the obstetric algorithms. We know this lineage firsthand: both of us served as editors of the Manual of Clinical Anesthesiology, one of the venues in which obstetric anesthesia crisis algorithms were developed and published. Those algorithms did not stay put. They evolved across editions and were ultimately carried into the standalone Stanford Obstetric Anesthesia Emergency Manual, a purpose-built cognitive aid for labor and delivery emergencies. Between venues and between editions, content changed. Steps were added, reordered, consolidated, and cut, because that is what responsible editors do as evidence accumulates. Institutions may also adapt cognitive aids to their own systems, resources, and workflows, creating another layer of variation between a published source document and the manual actually in use at the point of care.

The consequence is straightforward. The “same” checklist may exist in materially different versions, and the version hanging in an operating room at the time of an event may say something different from the version available for download today. Determining which version applied, at what time, and by what route it arrived at the institution calls for reconstructing provenance much like following a chain of custody. If an expert cannot explain the publication history of the algorithm in question, they are consulting the wrong edition of the truth.

**Time dependence, and why the interval matters more than the omission**

In our experience the most common analytic error in these cases, after the version problem, is treating a missed step as a binary event. Clinically it rarely is. Most crisis interventions are time dependent, and the relevant clinical question is almost never whether a drug was given, but how long it took to arrive.

A few of the interventions that appear in these manuals have a published literature relating outcome to time of administration, though the strength of that literature varies a great deal from drug to drug. Dantrolene in malignant hyperthermia is the clearest example, and it is genuinely quantified: the likelihood of a complication rose approximately 1.6 times for every 30-minute delay between the first sign of malignant hyperthermia and the first dose. (Larach et al., Clinical Presentation, Treatment, and Complications of Malignant Hyperthermia in North America from 1987 to 2006 (2010) 110 Anesth. & Analgesia 498, 505.) Epinephrine in anaphylaxis is real but weaker in kind – delayed administration is associated with fatal outcomes across retrospective series, but no study establishes a per-interval dose-response. (Pumphrey, Lessons for Management of Anaphylaxis from a Study of Fatal Reactions (2000) 30 Clinical & Experimental Allergy 1144.) Lipid emulsion in local anesthetic systemic toxicity does not belong on this list at all: no published study relates outcome to the interval before lipid administration, and the systematic review of the human literature graded the entire body of evidence as very low quality. (Hoegberg et al., Systematic Review of the Effect of Intravenous Lipid Emulsion Therapy for Local Anesthetic Toxicity (2016) 54 Clinical Toxicology 167.) When evidence of that kind is available, the variable that matters clinically is a time interval, and a record can support reconstruction of that interval with much greater confidence than a checklist can be used to infer a counterfactual.

This has a practical implication for anyone evaluating one of these cases. Reconstructing the interval, and situating it against what is published about that specific intervention, is the analysis most likely to produce a defensible opinion in either direction. It is also the analysis that most often shows a delay to be clinically immaterial, which is worth knowing early. Whether a reconstructed interval satisfies the legal test for causation is a question for counsel. What the clinical literature can say, often with surprising precision, is how outcome varies with time to treatment for the specific intervention at issue.

**Ten questions to ask about any anesthesia crisis record**

The questions that follow are those we consider before reaching a conclusion. They are medical and evidentiary in nature, not legal, and a number of them refer to proof that is often not asked for at all, including some items that may not be kept permanently. The first five concern the record itself, because a manual analysis built on an unexamined chronology will not hold; the last five concern the manual and the system around it.

1. What is the source and reliability of each timestamp? Anesthesia records mix machine-captured and hand-entered times, and a chronology built on unexamined timestamps is a chronology built on sand.
2. What do the vital-sign trends show, independent of the narrative note? The monitors tell their own story, and it does not always match the prose.
3. What was charted contemporaneously versus reconstructed afterward? Late entries are not suspicious by default, since charting stops when hands are busy, but the line between real-time and retrospective has to be drawn before anything is inferred from timing.
4. Does the monitor or anesthesia information system retain raw data beyond the printed record? Many systems hold granular physiologic data that never reaches the chart, and retention windows close.
5. What do the code cart log, automated medication dispensing records, and defibrillator memory show? Dispensing records independently corroborate or contradict a medication timeline, and modern defibrillators store downloadable, time-stamped event data including shocks delivered, pads-on time, and CPR quality metrics.
6. Which edition of the manual was in effect, and what is its publication lineage? Algorithms evolve across editions and venues, and the version question must be settled before anyone opines about specific steps.
7. When was the manual adopted, and was it accessible at the point of care? Adoption dates, committee minutes, and physical placement turn an abstract document into a concrete fact about that operating room. A binder in a workroom two doors away is a different fact from a binder at the workstation.
8. Is there documented simulation or crisis training for this department, covering which scenarios, attended by whom? Attendance rosters convert training from a policy into a fact about these clinicians, and they can support either view of the case.
9. Can crisis resource management roles be identified in the record, in testimony, or in communication logs? Leader, reader, help summoned. Paging and code-activation logs independently establish when help was requested and when it arrived.
10. Does the record suggest fixation or task saturation rather than a knowledge deficit? A team overwhelmed by simultaneous demands is a clinically different situation from a team that did not know what to do, and the two have different implications for everything downstream.

**What the reconstruction is for**

The work in one of these cases is not to hold up the manual. It is to reconstruct the crisis as it was lived: what the team knew, when they could have known it, what system the institution had built around them, what the cognitive aid could have contributed, and whether any interval that opened up was long enough to matter clinically.

Sometimes that reconstruction shows care that fell below what we would expect. Sometimes it shows an institution that adopted a safety system on paper and never implemented it. Sometimes it shows a capable team making 11 defensible decisions in a situation that was not winnable. We have reached all three conclusions in our own work, and we would be suspicious of anyone who reports reaching only one.

We say this as authors of these manuals. We built them because they help teams perform critical actions more reliably during crises: guiding a team through the worst four minutes of a patient’s life, strengthening the training that prepares that team, and reminding everyone in the room that even excellent clinicians need support under stress.

Mounted on the wall, the laminated binder was created to help a team save a life at 2:14 a.m. It was not intended to answer what occurred after that. Doing so calls for rebuilding the crisis itself: what the team knew, what actions they took, what support was available, how much time the critical interventions consumed, and whether any delay lasted long enough to be consequential.

*Larry F. Chu, MD, MS is professor of Anesthesiology at*Stanford University School of Medicine*, the founding co-author of the Stanford Emergency Manual and the Stanford Obstetric Anesthesia Emergency Manual, and co-editor of the Manual of Clinical Anesthesiology. He reviews anesthesia crisis and perioperative standard-of-care matters for plaintiff and defense counsel and can be reached at [This email address is being protected from spambots. You need JavaScript enabled to view it..](https://www.plaintiffmagazine.com/mailto:lchu@stanford.edu.)*

*Andrea J. Traynor, MD is professor of Anesthesiology and Perioperative Medicine at*Oregon Health & Science University School of Medicine*, where she serves as Anesthesiology Residency program director. She completed her anesthesiology residency and obstetric anesthesiology fellowship at Stanford University School of Medicine, lectures extensively on obstetric anesthesiology and crisis management, and is co-editor of the Manual of Clinical Anesthesiology. She reviews anesthesia crisis and perioperative standard-of-care matters, especially those concerning obstetric emergencies, for plaintiff and defense counsel and can be reached at [This email address is being protected from spambots. You need JavaScript enabled to view it..](https://www.plaintiffmagazine.com/mailto:andreatraynorconsulting@gmail.com.)*

### [Andrea J. Traynor](https://www.plaintiffmagazine.com/authors/andrea-j-traynor "Andrea J. Traynor")

[![Andrea J. Traynor](https://www.plaintiffmagazine.com/media/zoo/images/Traynor-Andrea_5e43c4b824defca188444ff71d36403e.jpg)](https://www.plaintiffmagazine.com/authors/andrea-j-traynor "Andrea J. Traynor")

*Andrea J. Traynor, MD is professor of Anesthesiology and Perioperative Medicine at*Oregon Health & Science University School of Medicine*, where she serves as Anesthesiology Residency program director. She completed her anesthesiology residency and obstetric anesthesiology fellowship at Stanford University School of Medicine, lectures extensively on obstetric anesthesiology and crisis management, and is co-editor of the Manual of Clinical Anesthesiology. She reviews anesthesia crisis and perioperative standard-of-care matters, especially those concerning obstetric emergencies, for plaintiff and defense counsel and can be reached at [andreatraynorconsulting@gmail.com.](https://www.plaintiffmagazine.com/mailto:andreatraynorconsulting@gmail.com.)*

### [Larry F. Chu](https://www.plaintiffmagazine.com/authors/larry-f-chu "Larry F. Chu")

[![Larry F. Chu](https://www.plaintiffmagazine.com/media/zoo/images/Chu-Larry_773fced95a7143291f095ca68cf67d5c.jpg)](https://www.plaintiffmagazine.com/authors/larry-f-chu "Larry F. Chu")

*Larry F. Chu, MD, MS is professor of Anesthesiology at*Stanford University School of Medicine*, the founding co-author of the Stanford Emergency Manual and the Stanford Obstetric Anesthesia Emergency Manual, and co-editor of the Manual of Clinical Anesthesiology. He reviews anesthesia crisis and perioperative standard-of-care matters for plaintiff and defense counsel and can be reached at [lchu@stanford.edu.](https://www.plaintiffmagazine.com/mailto:lchu@stanford.edu.)*

### Subject Matter Index

[Client Management & Expectations](https://www.plaintiffmagazine.com/tag/Client%20Management?Expectations=), [Law Office Management and Technology](https://www.plaintiffmagazine.com/tag/Law%20Office%20Management%20and%20Technology)
