Simulation Can Show Transfer, but It Cannot Inherit the Job

Thirty minutes in a simulation centre have become the phrase field ready.

A tab for every learner, in sim_assessment_results_FINAL_final2.xlsx, a pass column, and a dashboard you can show a board.

Efficient.

Also a small act of inheritance law that no simulation ever agreed to.

The scenario was a controlled representation.

The job is the job.

A simulation can show a decision under stated conditions. It cannot inherit the whole workplace.

That is not a complaint about simulation.

It is why simulation is useful.

The pass starts answering the wrong question

There is a real performance problem underneath.

A ward team calls for critical-care help late.

The on-call team arrives without a shared account of what has happened.

An educator and a clinical expert spend half an hour turning “improve communication” into two objectives.

Within five minutes of a defined deterioration, the bedside nurse states concern and requests senior review.

Before transfer, the team gives a structured handover: current status, relevant history, assessment, requested action.

Good.

That is a decision problem, not a slide-deck problem.

The simulation can make those decisions visible.

At minute three, the patient’s oxygen saturation and blood pressure change. At minute seven, a relative asks whether the patient is dying. At minute nine, a junior doctor receives another request.

At minute twelve, transfer and handover happen.

Each inject has a job.

Reveal escalation.

Prioritisation.

Delegation.

Handover.

The controller records facts.

Escalation was called at six minutes.

Blood pressure was absent from the call.

A task owner was named.

The family question went unanswered for ninety seconds.

Valuable evidence.

It is not evidence that the team will make every good decision on a crowded ward.

With the actual phone system.

A real family.

A senior colleague with a previous history.

And a patient who has declined to follow the timing sheet.

The evidence is bounded because the world was bounded.

That should feel unsurprising.

Instead, a pass marker is routinely asked to become a job description with a lanyard.

Start with the decision, not the room booking

Assessment starts with a claim, and the claim has to name a decision. What will this scenario result let somebody decide, and under which conditions?

“Completed emergency scenario” is an activity record, and you already have too many of those.

“Can recognise defined deterioration, request senior review, and deliver a structured handover within the specified scenario” is a claim somebody can inspect.

The difference is a great inconvenience to a dashboard, which prefers a single green dot.

But the claim tells you how much representation your programme actually needs.

If the target is recognising a protocol, a table-top or a controlled case may be enough.

If the target is coordinating an escalation through real roles, real information and real equipment, the conditions have to expose those things.

The World Health Organization distinguishes discussion-based tabletop exercises from operations-based drills, functional exercises and full-scale field exercises.

Those are not escalating badge levels.

They are different ways of making a decision visible while managing risk.

Your simulation needs the equivalent of an assessment blueprint.

A planned starting state.

Information flow.

Triggers after correct action.

Responses to omitted action.

A stopping point.

Otherwise it is a dramatic event with a smoke machine and no argument.

Fidelity is a condition, not a compliment

People say a simulator is high fidelity as though it has been promoted.

Fidelity is more useful treated as a statement about what the representation preserves, and what it does not.

Equipment orientation is not a polite preamble.

If a nurse cannot find the emergency call number, or a doctor cannot locate the observation chart, or a team cannot find the oxygen source, then later failure is ambiguous.

It may be judgement.

It may be that the room does not work like the work.

That is psychological fidelity doing operational work.

The fiction contract makes the boundary explicit.

Participants engage with the represented world.

Facilitators disclose material limitations, and do not turn those limitations into trick questions.

Ten minutes of prebrief for six participants can establish roles, equipment, the time limit, the stop phrase, recording boundaries, and what the manikin cannot do.

That is not ceremony.

It is part of whether an observed action can mean what the rubric later says it means.

Aviation makes the distinction unusually hard to ignore.

The FAA’s simulator framework separates flight training devices from full flight simulators across defined levels, with qualification attached to specific aircraft models, configurations, performance tests and permitted tasks.

A device level is not a generic ranking.

An unqualified virtual experience may be excellent practice.

It does not acquire regulatory credit because the headset looks like it has opinions about turbulence.

Every other sector merely has less obvious paperwork.

The scenario is doing real work, which is the difficulty

None of this argues for returning to a quiz with twenty questions and one suspiciously long answer option.

Simulation can safely expose people to rare, consequential events.

It can make an escalation decision, an information handoff, a prioritisation choice and a team’s division of work observable, in ways selected response cannot reach.

It can also reveal a system defect.

When the emergency call number is missing from a new ward phone display, the useful finding is a facilities request. Not a remedial course assigned to everyone who failed to discover it telepathically.

The healthcare simulation standards make the traceability principle explicit. A defined practice need connects to measurable objectives, to scenario events, to an observation method, to feedback, and to a decision.

The scenario, the rubric and the debrief are one production unit. Even when three teams own them and the folder is called Simulation_Assets_Archive_DO_NOT_MOVE.

Debriefing is what converts an observed episode into practice.

Twenty minutes of debrief first reconstructs the facts, then examines two or three objectives, then rehearses the replacement.

Name the escalation phrase.

Designate one caller.

Read the last two observation sets during handover.

Then, within 48 hours, the educator sends a de-identified summary.

Two strengths.

Two barriers.

Owners.

A due date.

That is how a staged event leads to a changed work condition.

The simulation is not pretending to be the job.

It is helping the job become safer and more observable.

A consequential score needs a chain of custody

The trouble starts when the result decides progression, credentialing or employment.

Then four controls matter.

Standardisation.

Rater preparation.

Scoring rules.

Appeal.

Standardisation fixes your starting state, your information, your role-player responses, your time window and your allowable prompts.

It does not promise an identical emotional experience.

It stops material differences being left to facilitator whim.

Raters apply the same rubric to sample performances and reconcile disagreement before any high-stakes use.

Every level has to name something a rater can see. “Good judgement” is an aspiration, not a scoring dimension.

Critical errors may be conjunctive: every specified safety criterion has to be met. That decision belongs in the scoring rule before candidates begin, not in a discussion afterwards.

So your assessment record needs the scenario version, the simulator configuration, the instructions, the allowed materials, the time window, the accommodation, the role-player script, the rubric, the scorer materials, the incident log and the result.

Rubric_2.1 explains a later review.

“Latest final revised” explains a family of avoidable meetings.

The same boundary protects the learners and the programme.

A candidate who misses an unfamiliar control may have a real deficiency.

If eight of ten experienced people miss it, the scenario or its promised orientation needs investigating.

A checklist does not settle that question.

Neither does a total score, however cleanly it fits into your HRIS.

Assessment does not convert a score into a permanent adjective.

It produces a qualified claim, with evidence and limitations attached.

The simulator log has met the talent platform

This is the current problem.

Simulation produces unusually attractive machine data.

Timestamps.

Clicks.

Decisions.

Prompts.

Video markers.

Checklist rows.

Rubric dimensions.

A reassuringly decimal score.

A model can ingest all of it beside learning history, job profiles and succession data, then infer that somebody who passed Scenario 3.2 has a generally portable capability.

The model is not being reckless.

Your data contract is being coy.

Scenario data needs the same provenance as any other workforce evidence.

With extra care, because it looks so complete.

Keep the decision claim.

The scenario version.

The equipment and configuration.

The role assignment.

The inject sequence.

The administration conditions.

The rubric version.

The rater record.

The incident record.

The score.

The limitation.

Let an AI system propose a learning action, or flag a question.

Do not let it silently promote a scenario-bound observation into a universal job fact.

Your analytics layer thinks timestamps are facts.

They are facts about a controlled event.

Validate transfer where the work happens

The durable object is not a simulation pass.

It is the ability to recognise the condition, choose the action, verify it, escalate it and adapt, in the real operational setting.

Which is harder to store.

Naturally your system would prefer the pass.

So your transfer claim needs a second evidence path.

After the session, record participation, objectives, observations, corrective actions and completion.

Test the operational change with a three-minute micro-drill where that makes sense.

Repeat a scenario after the point of prior success, and change the conditions, so the group has to apply a principle rather than recite a script.

Then observe the target behaviour in the field, and inspect whether your work environment now supports it.

The published evidence base is honest about its own limits. A systematic review of crisis-resource-management simulation found nine eligible studies.

Four measured workplace transfer.

Five measured patient outcomes.

The finding was suggestive rather than conclusive.

That is not disappointing.

That is the honest shape of the claim.

A simulation can show that a team recognised a cue, selected an action and coordinated a response inside a designed world. It can surface an instructional need, or a system barrier that deserves attention on Monday morning.

It cannot absorb the equipment, the pressure, the consequences, the local workarounds, the team history and the genuine uncertainty of the job, because the result field says PASS.

Keep the pass.

Keep its boundaries with it.