Drownproofing: Can You Teach Resilience?

At EMS World Expo, “Drownproofing: Can You Teach Resilience?” will examine how to design stress exposure that strengthens performance instead of merely measuring failure, and how Systems 1, 2, and 3 are changing preparation for the call that does not follow the script.

They say, “Go Big”

We thought we had it nailed.

When I became operations chief, our department here in rural Alaska was going through a period of significant evolution. One of the pillars we knew we wanted to build out on was our internal training program. Building from the ground up, we had the backing of our leadership and the resources to do it.

We had scheduled training, mandatory attendance, medical director involvement, and enough content to keep everyone occupied until retirement. We taught the algorithms, practiced the procedures, and checked all the boxes.

So, we went big.

The following summer we held a large-scale, multi-agency mass casualty exercise requiring triage, prioritization, resource management, communication, and scene control. It looked excellent in the photographs.

And we failed.

Despite all our preparation, the didactic and skill stations, and all the classwork we had covered, each group missed on significant objectives.

How could this happen?

We went back to the drawing board, took the failed objectives and developed training specific to them. Designed another multi-agency drill with another setting. Did all the preparation. (The photos are even better for this one, because there are dogs in them.)

Mass casualty in a remote setting is what we need to be good at. It is not optional.

And we failed again. Same exact issues, one year later.

Not because our providers were unintelligent, indifferent, or untrained.

They knew the material; they simply could not reliably retrieve, organize, and apply it once the water began to rise.

EMS education is good at delivering information and testing whether someone can reproduce it under controlled conditions. It is far less consistent at teaching people to think when the scene is loud, information is incomplete, the patient is deteriorating, and every option carries a consequence.

We keep adding content when the actual problem may be performance under cognitive load.

Stress Is Not an Instructional Design Strategy

In EMS, we sometimes confuse difficulty with rigor. We make scenarios louder, darker, bloodier, and more chaotic, then congratulate ourselves for creating "realism.”

But merely stressing a learner does not create resilience. It can overwhelm working memory so thoroughly that the learner's practice is freezing, guessing, or waiting for rescue.

Recent work on stress during paramedic and healthcare simulation reaches a practical conclusion: some stress improves engagement and realism, but beyond a threshold, it impairs learning and performance. The challenge is not to remove stress; it's to dose it deliberately.[i]

That is the idea behind “drownproofing.”

Traditional drownproofing doesn't teach a swimmer to overpower the water. It teaches the person to conserve effort, regain position, control breathing, and return to a sequence that works. In EMS training, it means building enough automaticity, pattern recognition, and recovery skill to keep functioning when conscious reasoning becomes expensive.

Resilience in this context isn't stoicism or emotional invulnerability, nor is it an excuse to expose people to preventable harm and call it character building.

The Two Systems Already in the Ambulance

Daniel Kahneman popularized the language of System 1 and System 2 thinking in his book, “Thinking Fast and Slow.”[ii]

System 1 is fast. It recognizes patterns, reaches for familiar scripts, and produces an answer with little conscious effort. In experienced clinicians, it notices that the patient "looks wrong" before the monitor explains why. It's also where heuristics and bias live. Prehospital critical care literature identifies anchoring, availability, confirmation, framing, overconfidence, premature closure, and omission among the biases that can shape high-stakes decisions.[iii]

System 2 is slower and deliberate. It compares possibilities, questions first impressions, and works through unfamiliar problems. It is effortful and dependent on limited working memory resources.

Under pressure, System 2 becomes harder to access cleanly. Acute stress can impair working memory and cognitive flexibility, while the physical, environmental, and informational demands of prehospital care can drive cognitive load in either direction.[iv] [v]

The answer isn't to choose one system over the other. The goal is to build a dependable System 1 foundation that creates room for System 2. Repetition can make equipment location, initial assessments, role assignments, and high-risk procedures less cognitively expensive, preserving bandwidth for the part of the call that doesn't match the algorithm.

This is precisely what makes practices like well-rehearsed “Pit Crew CPR” so successful.

Introducing … System Number Three

There is now another influence in the ambulance: external artificial cognition.

A recent Wharton preprint describes artificial intelligence as "System 3," external reasoning that can influence both intuitive and deliberate judgment. The researchers call it cognitive surrender when a person adopts the system's answer with insufficient scrutiny.[vi]

EMS has always used cognitive aids: protocols, checklists, calculators, decision rules, and consultation can reduce error. The danger is allowing the tool to replace the mental model needed when it is unavailable, incomplete, or wrong.

Training must preserve productive struggle. Before reaching for an answer, the learner should commit to an assessment, name the relevant cues, and explain the intended action. The protocol or AI tool can then challenge, verify, or refine that reasoning.

Used this way, System 3 supports cognition.

Used answer-first, it may quietly prevent System 1 and System 2 from developing at all.

Stop Throwing Everyone into the Deep End

Our failed mass casualty exercises taught us that scale isn't the same as sophistication.

We'd combined too many unstable skills at once, expecting learners to manage communication, triage, command, treatment, transport priorities, unfamiliar partners, and environmental confusion before components were durable on their own.

Since then, we’ve gone smaller.

We narrowed the objectives, used small groups, repeated short scenarios, introduced unpredictable responses, and gave immediate, specific feedback. We returned to the skills after time had passed rather than assuming one exposure created permanent competence.

Only then did we add stress.

Good stress exposure training is graduated. Change one variable at a time:

• noise
• time pressure
• equipment failure
• a distraught family member
• an incomplete handoff
• a second patient or partner who needs direction

The learner should be challenged but still able to process feedback.

Adaptive stress training research supports this principle. In a 2025 study, stress exposure adjusted to the learner's real-time response produced strong results on several stress and engagement measures than a fixed progression. The lesson does not require virtual reality. Training should respond to the learner, not the scenario schedule.[vii]

A brief mental skills lesson is also unlikely to produce durable change alone. In one randomized emergency medicine study, a short didactic intervention didn't significantly change stress measures. The authors pointed toward longitudinal reinforcement and deliberate practice.[viii]

The debrief must examine the decision process, not merely the final answer. What did you notice first? What did you miss? When did the scene stop making sense? What cue changed your plan? Did you decide or wait for someone else to decide? What would help you recover faster next time?

That is where resilience is built, not in the adrenaline, but in the return.

When The Hands Know Before the Brain Does

Procedural memory stores the learned sequences used to perform a skill. It allows an experienced clinician to spike a bag, assemble an airway, apply a tourniquet, or move through an assessment without consciously narrating every step.[ix]

Early in learning a skill, it depends heavily on working memory. The learner must remember what comes next, locate equipment, monitor hand placement, and compare each action against instruction. This is cognitively expensive.

With correct repetition, performance becomes more stable and automatic. Neural control is distributed rather than housed in one neat center, but the basal ganglia, cerebellum, motor systems, and their connections contribute to learning and refining practice sequences.[x]

The clinician doesn't stop thinking. The predictable portion of the task simply consumes less attention, preserving working memory capacity for what is changing.

That matters because stress can impair working memory, attention control, and cognitive flexibility. A well-established sequence may remain more accessible when those higher-order functions are strained.

Automatic behavior isn't automatically correct. Procedural memory preserves poor technique, shortcuts, and faulty sequencing as readily as good practice. Under pressure, people often revert to what they rehearsed most, not what they were taught most recently.

The goal isn't mindless repetition. It's to make the predictable elements automatic so the clinician can recognize when the situation is no longer predictable.

A Basic Skill-Layering Drill

Take any skill and break it down into its simplest components. Then remove any subjective interference from its use, take away the “noise.” There is no scenario here or use conditions on it (at first).

• Build a clean sequence. Practice slowly with the correct equipment. Accuracy comes before speed.
• Repeat until attention is freed. Continue until the mechanics no longer consume all concentration or require any additional prompting.
• Add ONE complication at a time. Use noise, limited space, missing equipment, changing vital signs, family member or partner questions, but not all at once.
• Force a decision AND a recovery. Recognize when the sequence no longer fits. State the problem. Change the plan and return to an organized process. Then debrief the cue.

Confirm that the learner recognized when the pattern changed and were able to anticipate and correct the problem.

A skill like ventilation would be easy to superimpose onto this drill.

1. Begin with positioning, mask seal rate, volume, and chest rise feedback.
2. Add pulse assessment or monitor interpretation.
3. Introduce a leak, worsening compliance, a poorly positioned airway, or a partner ventilating too quickly.
4. The learner must recognize that the familiar sequence is no longer working correctly and resume organized care.

Bagging is a foundational skill with life-threatening consequences if done poorly or incorrectly. It requires large amounts of focus yet isn't taught to a degree of automatic competency. This is a drill that can be done with any provider level. Layered with other skill practice, drills like this can raise the competence of an entire team.

What these repetitions do is move more reactions from System 2 over to System 1, without the need to force them into an unpredictable crisis to do it.

Only one of the learning objectives here is correct ventilation technique. The other, and more important objective, is did enough of that procedure become automatic so that the learner now has a little extra room to focus on the million other things that will be occurring during a crash airway in real time.

Competence Buys Time