guide
EMT Scene Size-up and Safety study guide
The six scene size-up and safety job tasks in the order a call meets them — a plan of action before arrival, hazards and who owns them, PPE chosen from the hazard assessment, investigating the scene for patients, triage in a published system, and the resources request — each with a worked scenario, plus the hazard list, the four PPE rules OSHA states, what universal precautions actually requires, and the failure modes that cost items on the first domain of the exam.
Scene Size-up and Safety — the first five minutes decide the call
What this topic actually asks of you
Scene Size-up and Safety is 15–19% of the cognitive examination, and it is the first of the five content domains — the one the other four assume is finished. The Registry prints its six job tasks in the order a call meets them, and the whole domain fits in six lines:
- Develop a plan of action prior to arrival using all available information to ensure safe and effective operations.
- Protect self, other responders, patient, public, and the emergency scene from existing and potential hazards.
- Don appropriate personal protective equipment (PPE) based on known hazards to protect responders from exposure.
- Investigate the scene to determine potential patients.
- Triage patients to ensure optimal care.
- Request appropriate resources based on known hazards and patient conditions to mitigate all facets of the emergency scene.
Six tasks out of the exam's 28, and every one of them happens before a treatment is started. That is the shape of the domain: it is small, it is early, and it is the part of the call that a candidate gets wrong on the exam for the same reason they get it wrong on the street — they treat it as paperwork to get through rather than a decision to be made.
The six tasks are one loop, not a checklist. A plan of action is a hypothesis, and the scene answers it: the plan said the front door, the front door is blocked by a car; the plan said one patient, there are four. What matters is what you do with the answer. The plan is rewritten, the hazards are re-checked, the equipment is chosen again for what was actually found, the patients are counted, sorted, and the resources that the real scene needs are requested — not the ones the dispatch imagined.
The rule that holds the domain together. The scene you were sent to and the scene you arrive at are two different scenes, and the second one is the one that can hurt you. Nothing in the primary assessment is worth arriving at an unsafe scene to start.
Where the numbers stop. This is the domain that most tempts a guide into printing a figure — how far back to park, how long a suit buys you, which exposure limit applies — and every one of those numbers belongs to a chemical, a service's policy or a local mass-casualty plan, where it carries a date and a scope. What the exam asks is which document owns the number, not what the number is.
1. A plan of action before you get there
Dispatch gives you less than you want and more than you use. Read it as a hypothesis: the call type, the address, the time of day, the floor, the number of callers, the words the caller chose, and anything the dispatcher added — a hazard flag, a history note, a note that the last crew was refused.
The plan itself is four decisions made out loud, in the ambulance, before the wheels stop:
- The approach — which direction to come from, where to leave the vehicle, and what that spot protects you from. Parking is a decision about traffic, fire, glass and escape, not about parking.
- The access — which door, which stairwell, which floor, and what you will do if that route is the one that is blocked.
- The roles — who carries what, who talks to the patient, who handles the airway bag, and who is watching the scene while the others work.
- The exit — the route out with a patient on a stretcher, and the point at which you stop, back out and call for something bigger.
A plan is not a prediction. The value of saying it out loud is that the crew can compare it with what is really there and say "not this" fast. A plan nobody says out loud cannot be revised, because nobody knows what is being revised.
What dispatch cannot tell you, and you should still ask. How many people are on scene and whether any of them are the reason the call was made; whether the address is one the crew has been to before; whether there is a hazard the caller did not mention because the caller does not know it is a hazard; and whether the nearest hospital is the right hospital for what the call is turning out to be.
Worked example — the third floor with no elevator
Dispatch: a 62-year-old man, "can't catch his breath", third floor, no elevator, one caller who has already hung up.
- 1 — Approach. Note the building is on a one-way street with a bus lane; the plan is to put the vehicle past the door rather than in the lane, so the stretcher has a straight run out.
- 2 — Access. Front stairwell, and the plan already names the rear stairwell as the alternative, because a third-floor carry-out with one usable stairwell is a plan with no margin.
- 3 — Roles. One EMT with the monitor and the airway bag, one with the stretcher bag, and the driver stays with the vehicle to move it if the street needs clearing.
- 4 — Exit, and the threshold. If the patient cannot sit forward without distress, the plan stops being "walk him down" and becomes "call for a stair chair and a second crew" — decided before the door opens, not at the top of the stairs.
2. Protect yourself first — and name what you are protecting everyone from
Task 2 has an order inside it: self, other responders, patient, public, and the emergency scene. The order is not arrogance. A responder who is inside the hazard becomes a second patient and removes the crew that was meant to help the first — which is the EPA's point in its own words:
"Always put your own safety first."
"Non-professionals who attempt to respond to these incidents may become part of the problem, not part of the solution."
The hazard list is longer than "fire". Work it out loud, because a hazard nobody named is a hazard nobody is watching:
- Traffic — the scene where a car comes through the work area is the classic second casualty.
- Fire and smoke — the fire is visible and the products of combustion are the reason crews wear equipment for the air, not for the flames.
- Chemicals — an unknown odor, an unlabeled container, a vapor cloud, a leaking drum, a garage with a smell in it and someone asleep in a chair.
- Electricity — a downed line, a vehicle against a pole, a battery, a solar array, water with a conductor in it.
- Structure and vehicle stability — a damaged building, a sagging stair, a car on its side, a load.
- People — a patient with a weapon, a crowd, a scene where the assault is still in progress, a domestic call where the person who asks you in is not the person you should be watching.
- Animals — a dog that is doing its job, even when you are doing yours.
- Environment — heat, cold, water, ice, a slope, a distance from the road.
- Blood and body fluids — the exposure you take with you (§3).
- Other responders — the ladder truck reversing, the engine moving, the scene nobody is directing.
Most of that list is not yours to fix. A fire, a chemical release, a downed line, an armed person and a hostile crowd are made safe by fire, hazmat, the utility and law enforcement; your job is to stop at the boundary, work where you can be protected, and call for the resource that owns the hazard (task 6). The name for that discipline is staging: the crew waits where it is safe, asks for what the scene needs, and treats only when the scene can hold a treatment.
A hazard flag is not a hazard assessment. A dispatcher who says "possible chemical" gives you a reason to park upwind and call; the assessment that decides anything is made by looking, and it changes as you get closer. Which distance, which zone, which protective action and which suit belong to the chemical itself and to the Emergency Response Guidebook; a distance copied out of context can move a responder into the hazard instead of out of it.
Worked example — the garage with a smell in it
Dispatch: "male down in the garage, wife says there is a strong smell, she is outside and she is coughing."
- 1 — Stop at the door. One patient is down and one caller is coughing: that is a pattern, not two complaints, and it says the problem is in the air of the space, not in the man.
- 2 — Move the caller out and keep her out — she is the patient you can already talk to, and she is the reason the scene needs no second rescuer inside.
- 3 — Ask for the resource that owns the hazard: fire and hazmat for the air and the atmosphere, the gas utility if the smell reads as gas, and additional ambulance capacity for the patient the crew cannot yet reach (§6).
- 4 — Treat what is in front of you — the patient who is out of the garage gets oxygen and a reassessment; the patient who is inside waits for the crew with the equipment for the space he is in. That is not neglect, it is the only version of the rescue that ends with a living rescuer.
3. PPE that fits the hazard you found
Task 3 is the shortest task and the one with the most rules under it, because three of them are law. Start with the rule that makes the rest make sense:
"Universal precautions shall be observed to prevent contact with blood or other potentially infectious materials. Under circumstances in which differentiation between body fluid types is difficult or impossible, all body fluids shall be considered potentially infectious materials." — OSHA,
1910.1030(d)(1)
The course uses the standard's own word, universal precautions, rather than the broader label services use for the same idea. The consequence is the same either way: the decision to wear gloves is not made from a diagnosis, because at the door there is no diagnosis — there is a patient and a fluid you cannot identify.
"Appropriate" is a test, not a mood. The standard defines the word:
"Personal protective equipment will be considered "appropriate" only if it does not permit blood or other potentially infectious materials to pass through to or reach the employee's work clothes, street clothes, undergarments, skin, eyes, mouth, or other mucous membranes under normal conditions of use and for the duration of time which the protective equipment will be used." — OSHA,
1910.1030(d)(3)(i)
Read the second half twice. Gloves that tear, a gown that soaks through, or an eye shield that leaves the mouth open fails the word "appropriate" — the equipment has to hold for as long as the equipment is being used, not for the first minute of it.
The equipment is a list, and it is the standard's list: "eyes, face, head, and extremities, protective
clothing, respiratory devices, and protective shields and barriers" (OSHA, 1910.132(a)). Match the list
to what the assessment found, not to the call type: a scheduled transfer does not need what a hazardous
scene needs, and a bleeding patient does not need respiratory protection.
The four rules that survive the exam and the street:
- Choose from the hazard, not the habit. The assessment of what the hazards are comes first and the
equipment follows it — "The employer shall assess the workplace to determine if hazards are present, or
are likely to be present, which necessitate the use of personal protective equipment (PPE)" (OSHA,
1910.132(d)(1)). - Fit. "Select PPE that properly fits each affected employee" (OSHA,
1910.132(d)(1)(iii)). A mask that leaks at the edge, or gloves three sizes out, are equipment that is not doing its job. - Damaged equipment is not equipment. "Defective or damaged personal protective equipment shall not
be used" (OSHA,
1910.132(e)). A creased face shield or a torn glove is replaced, not worn. - Off before you leave, and hands after that. "All personal protective equipment shall be removed
prior to leaving the work area" (OSHA,
1910.1030(d)(3)(vii)), and "Employers shall ensure that employees wash their hands immediately or as soon as feasible after removal of gloves or other personal protective equipment" (OSHA,1910.1030(d)(2)(v)). The two halves are one instruction: the equipment keeps the contamination on the equipment, and the handwash ends the exposure.
Don and doff is a trained sequence, not a guess. The standard requires an employer to train employees
in "how to properly don, doff, adjust, and wear PPE" (OSHA, 1910.132(f)(1)(iii)) precisely because the
sequence is where contamination is transferred — the outside of a glove touching an inner wrist, a mask
pulled down to talk. Learn the sequence your service trains, and practice the removal as often as the
wearing.
What PPE is not. It is not scene control. Gloves are not a reason to reach into a hazard, gloves are not a handwash, and a mask does not make an unsafe scene safe. The order is the guidance's own, and it puts the cheapest control first: "Proper hand washing is the most effective way to prevent the spread of infections in hospitals" (MedlinePlus, Infection Control).
Worked example — the cut hand on a kitchen floor
Dispatch: "40-year-old, cut hand, bleeding is controlled, patient is sitting on the floor."
- 1 — Gloves on before the door, and a second pair in the pocket, because the exposed hand is the one you will be holding while the first pair does the work.
- 2 — Look before you touch. Blood on the floor is a slip hazard, a shoelace, and a surface the patient's other hand has already touched: the hazard list in §2 is being worked while the gloves go on.
- 3 — Choose for the finding. A cut whose bleeding is already controlled calls for gloves; eye protection is for the call where a splash, spray or spatter can reasonably be expected, and this one is not. Adding equipment the finding does not call for is its own kind of error.
- 4 — Doff before leaving the work area, then wash. The equipment comes off at the kitchen door, the hands are washed before the stretcher is touched, and a torn pair is replaced rather than worn.
4. Investigating the scene to find the patients
Task 4 is "investigate the scene to determine potential patients" — and the operative word is potential. The patient in front of you is the easy half. The other half is the patient who is not in front of you yet: the passenger in the back, the child in the seat, the person who walked out of the building and sat on the curb, the second occupant of the apartment, the worker who went home before the ambulance arrived.
Start with the mechanism of injury or the nature of illness. Every call is one or the other. A mechanism of injury is how energy reached the body — the fall, the collision, the weapon, the blast, the electricity — and it tells you what to look for even when the patient looks fine. A nature of illness is the medical complaint the patient can describe, and it tells you what the scene should look like: a diabetic emergency and its juice carton, a respiratory call and its inhaler, a seizure and the evidence on the floor.
The clues are in the room, and they are read, not searched. Medicines on a table, a pill organizer, an oxygen concentrator, a hospital band, a walker, a note, a phone, a needle, a weapon, a dog, a stairway with no rail, a carpet with a turned corner, a seat belt still buckled across an empty seat: each is a sentence about the patient and about the scene. Reading them is scene work; opening drawers is not.
Count, then say the number out loud. The count decides everything downstream: one patient is a call, four patients is a triage (task 5) and a resource request (task 6). The count is also the number most often wrong on the first pass, because the walking wounded leave the scene, the quiet patient stays, and nobody counts the people standing up and saying they are fine.
An index of suspicion is not a diagnosis. The finding that matters most on a scene is the one that does not fit: the damage that looks minor and the patient who cannot remember the last minute; the apartment that smells of gas and the patient who says he is only tired. Investigating is deciding how much more scene to look at before you touch a patient, not deciding what is wrong with them.
Worked example — the car off the road
Dispatch: "single vehicle off the road, driver out and walking, no other information."
- 1 — Look at the car before the driver. One door open, one door shut, a child seat in the back with the harness still buckled, and a rear window that is gone: the car is telling you the count is wrong.
- 2 — Ask the driver where everyone else is, as a question with a number in it: how many people were in the car, and did anyone get out and walk? A patient with a head injury is a poor witness and a good source of the count you cannot take yourself.
- 3 — Walk the perimeter before you load. The person who got out and sat behind the barrier is still a patient, and the count of patients is what the resource request in §6 is built on.
- 4 — Investigate the mechanism, not the complaint. "Driver out and walking" is a mechanism finding — a car that left the road and lost a window — and a mechanism that large is a reason to look at the driver again rather than to accept the driver's own report of how he feels.
5. Triage — sorting patients in a published system
Triage is sorting, not treating. It answers one question per patient — how soon does this one need to leave — and it exists for the scene where there are more patients than there are crews. A single patient is not triaged; the patient is assessed and transported. The moment a scene has more patients than the crews can reach, somebody has to decide the order, and that decision is triage.
START is one published system, and it is one system, not the rule. The algorithm's own provenance:
"START was developed by the Newport Beach Fire and Marine Department and Hoag Hospital in Newport Beach, California in 1983."
"At present START remains the most commonly used mass casualty triage algorithm in the US."
Its five parameters: the ability to walk; the presence or absence of spontaneous breathing; a respiratory rate on either side of a published figure; perfusion, read as a palpable radial pulse or visible capillary refill; and mental status, read as the ability to obey commands. Four labels come out of it:
| Label | Tag | What the tag means |
|---|---|---|
| Minor | Green | Relatively minor injuries; may be able to help with their own care — often called the "walking wounded" |
| Delayed | Yellow | Transport can be delayed; the status is not expected to worsen significantly over several hours |
| Immediate | Red | Can be helped by immediate intervention and transport; needs medical attention within minutes for survival |
| Expectant | Black | Unlikely to survive given the severity of injuries, the available care, or both |
The first parameter is the cheapest and the one candidates forget. The ability to walk sorts a group faster than any examination, because it needs no equipment and no hands. The patients who walk are not the patients who get the ambulance first — that is the whole point of the sort.
A tag is a decision recorded, not a diagnosis made. What the algorithm supplies is the sort itself — the parameters, the label each combination produces, and the color that goes with the label. What an agency writes on the card, and where the patient goes once it is tied, belongs to the local mass-casualty plan, which also owns the treatment area, the transport priority and who is in command of the scene. START is a published sorting method that a plan adopts, and the plan is what a service trains to.
Re-triage, because sorting changes. A patient who was breathing on the first pass may not be on the second, and a patient tagged delayed who is now pale is a patient whose tag is out of date. The second pass is cheap; the wrong tag is expensive.
Worked example — four patients outside a two-car collision
Findings: one adult walking and holding a phone, one adult sitting on the ground talking but pale, one adult lying still and breathing, one adult with an open leg wound who is awake and asking for help.
- 1 — Sort the walking first. The adult on the phone is the minor of the group: green, directed to the collection point, and told to stay there — and counted, because a patient who is directed away is still a patient on the board.
- 2 — Then the ones who are not walking and not breathing normally. The pale adult is a perfusion finding; the adult lying still needs the breathing and the mental-status parameters read before a tag.
- 3 — Say the labels out loud to the crew, with the count: four patients, one green, and the others sorted by the parameters rather than by who is shouting. The loudest patient is not the first patient.
- 4 — Call the resources the count implies (§6): one ambulance cannot move four patients, and the request goes out on the first pass, not after the last tag is tied.
6. Requesting the resources the scene needs
The sixth task is the one that turns a scene from a problem the crew is losing into a problem with a plan: "Request appropriate resources based on known hazards and patient conditions." Both halves matter. The hazards say what kind of resource (fire, hazmat, utility, law enforcement, rescue); the patient conditions say how much of it, and how fast.
What an EMT can ask for, and what it is for:
- More ambulances — because the count in task 4 was higher than the number of stretchers on scene.
- Advanced life support — because a patient's condition needs an intervention the crew cannot provide.
- Fire and rescue, with extrication — because the patient cannot be reached or cannot be freed without tools and a crew whose job that is.
- Hazardous materials — because the hazard is in the air, in a container, or on the patient, and it is not yours to control.
- Law enforcement — because the scene is not safe from a person, or because traffic needs stopping.
- The utility — because the wire is down, the gas is leaking, or the water is carrying current.
- Air medical — because the distance or the patient's condition makes the ground route the slower one; the local plan says who may request it.
- A mass-casualty response — because the count crossed from "several" into "more than this system can move with what is here".
- The medical examiner or coroner — because a patient is dead and the scene is now that agency's; the local plan and state law say who calls and who pronounces.
Ask early, and ask in a sentence that can be acted on. Resources take time to arrive, and the request costs seconds. The message is the same shape every time: what you need, where it is needed, how many, and what for — "two additional ambulances to the corner of Fifth and Main for four patients from a two-car collision, one of them critical" is a request; "we could use some help" is not.
A chemical release is reported, not just worked. The EPA's own instruction:
"Report hazardous materials releases and oil spills by dialing 911, or call the National Response Center at 1-800-424-8802."
Asking is a decision, not an admission. The crew that calls for help on the first pass is the crew that has time to treat the patients it already has while the help is coming. The crew that waits to be sure spends the whole scene behind the scene, and the resource nobody asked for is the one that arrives after the patient needed it.
Do not do another agency's job while waiting for them. Holding the boundary of a hazard, keeping traffic off a scene and refusing to enter an atmosphere are all things an EMT does with the resource on the way, not instead of it.
Worked example — three pinned in a two-car collision
Dispatch: "two-car collision, one lane blocked, three people reported trapped, traffic is moving past on the shoulder."
- 1 — The hazards set the first two requests: fire and rescue for the extrication tools and the traffic, and law enforcement for the lane, because the moving traffic is a hazard to the patients, the crews and the people already out of the cars.
- 2 — The conditions set the rest: three trapped patients is three ambulances, and if one of them is not answering, the request includes advanced life support rather than waiting to see how the extrication goes.
- 3 — Say where and what for. The staging point is named, the approach is named, and the request says which patient the first ambulance is for — a request that a dispatcher can read back is a request that arrives.
- 4 — Keep the request open. A count that turns out to be four, or a patient who deteriorates while the roof is being cut, is a second request; the first one was answered for the scene it was made on.
The failure modes, and the fix for each
Every failure below is a way of having the six tasks and not having the domain. They are listed in the order they cost points.
| Failure | What it looks like | The fix |
|---|---|---|
| Treating the dispatch as the scene | Walking in on the address, not on what is actually there | Say the plan out loud, then compare it with the scene |
| Scene safety as a formality | "Scene safe" said to a partner, not to the hazard | Name every hazard out loud, in the list in §2 |
| Parking as parking | The vehicle left where it is convenient | Park for the traffic, the escape and the stretcher |
| PPE from habit | Gloves for a scheduled transfer, gloves for a chemical cloud | Choose from the hazard assessment, not the call type |
| The glove as a handwash | Gloves off, stretcher touched, hands never washed | Doff before leaving the work area, then wash |
| Damaged equipment worn | A torn glove kept on because the patient is bleeding | Replace it: the standard says defective equipment is not used |
| The visible patient as the count | One patient loaded, one passenger still in the car | Look at the vehicle, the building and the walking wounded |
| Triage by volume | The loudest patient tagged first | Sort by the parameters, and count the walking |
| A tag never revisited | A delayed patient who has gone pale, still tagged yellow | Re-triage on every pass |
| Waiting to be certain before asking | The request made after the scene is out of hand | Request on the first pass, in a sentence that can be acted on |
How to answer the questions
This domain is 15–19% of the cognitive examination, and a question here comes the way the domain works: a scene, an incomplete picture, and a question about what the EMT does next or what a finding means. It does not ask you to name a hazard in the abstract; it puts an EMT in a doorway with something wrong on the other side.
Read the scenario for the task it is standing on. "A crew arrives to find a patient collapsed in a garage with a strong chemical odor and a second person coughing at the door — what should the crew do first?" has one answer among the six tasks: the hazard assessment before the patient. A candidate who starts the primary assessment has answered a different question.
Two of the six tasks are recall and four are judgment. The recall ones are set pieces — what the PPE list covers, what the four triage labels mean, what universal precautions require. The judgment ones are the reason this domain is worth 15–19%: given this scene, what is the next action, and what does the finding in front of you mean?
Every claim about the scene is a claim you can defend. "Scene safe" is not a sentence; "traffic is stopped, the wire is down and the utility has been called, and the atmosphere is the reason we are not going in" is. The answers that sound like safety and are not: parking where it is convenient, donning the equipment the call type suggests, tagging the patient who is loudest.
Where the numbers stop. No isolation distance, exposure time, exposure limit, decontamination recipe or PPE level is a fact to memorize, because each belongs to the chemical, the service's policy and the local plan, where it carries a date and a scope. The six tasks, and the order they run in, are what the domain is.
How to practice this
- Work the free scene size-up and safety practice set one item at a time, and before reading the choices, name which of the six tasks the scenario is standing on. The task decides the action; the facts only say which task it is.
- Rebuild the loop from memory each session: plan → hazards → PPE → count the patients → sort them → request what the scene needs. Then say the rule that moves it: the scene you arrive at is not the scene you were sent to.
- Take the hazard list in §2 and walk one real room with it — your own kitchen, a stairwell, a parking garage — and name every item on the list you can find. The list is only useful if it is automatic.
- For each worked example, cover the bullets and answer the patient yourself. The findings are written to be enough; if you have to guess, the finding you missed is the one to reread.
- Every item in this practice set is multiple choice. The real exam also uses multiple-response, options-table, build-list and drag-and-drop items — all scored all-or-nothing — and you should look at the official samples before test day.
Where to go next
- What happens after the scene is safe — the Primary Assessment is the domain this one hands over to, and it is the largest single share of the exam.
- The history and the focused exam — Secondary Assessment is what follows the primary assessment, once the patient is stable enough to be asked.
- Where the treatments live — Patient Treatment and Transport is the domain that owns the doses, the rates and the skills this one deliberately does not teach.
This topic is 15–19% of the cognitive examination — the Registry's own published range for the Scene Size-up and Safety content domain, and the first of the five.
Independent study material for the National Registry EMT Certification Examination. Elk Ridge Academy is not affiliated with, endorsed by, or sponsored by the National Registry of Emergency Medical Technicians. "NREMT" and "National Registry of Emergency Medical Technicians" are trademarks of the National Registry of Emergency Medical Technicians.
Source: NREMT — EMT Examination Specifications; OSHA 29 CFR 1910.132 and 1910.1030; REMM START Adult Triage Algorithm; US EPA — emergency response and hazardous substance releases; MedlinePlus — Infection Control (accessed 2026-11-08 and 2026-11-10)