The most dangerous object in your house
Ranked by how many people it sends to the emergency room, the most dangerous object in an American home is the bed. Ranked by the hour, it is one of the safest things you own. Here is what happens when you divide by time.
In 2024 the bed put an estimated 463,744 Americans in an emergency room. That is more than the stairs, more than the bathtub, more than every kitchen knife in the country. By raw count it is the single most dangerous object in the house.
It is also where you spend nine hours a day.
Every “most dangerous household object” list is really a list of things you are near a lot. The interesting question is not how many people an object hurts — it is how many it hurts per hour of exposure. So: national ER-visit estimates for 2024, divided by how long Americans actually spend doing the thing each object belongs to.
Dividing by time does not reorder the list. It inverts it.
| object | ER visits 2024 | by count | per million hours | by rate |
|---|---|---|---|---|
| Bed | 463,744 | 1 | 0.52 | 7 |
| Stairs | 460,492 | 2 | 46.28 | 1 |
| Bathtub / shower | 248,641 | 3 | 3.73 | 3 |
| Kitchen knife | 164,716 | 4 | 2.33 | 6 |
| Sofa / couch | 118,529 | 5 | 0.46 | 8 |
| Ladder | 75,816 | 6 | 6.93 | 2 |
| Lawn mower | 60,902 | 7 | 3.22 | 4 |
| Power saw | 32,551 | 8 | 2.97 | 5 |
| Stove / oven | 23,758 | 9 | 0.34 | 9 |
| Trampoline | 3,780 | 10 | 0.10 | 10 |
The stove — an object that is on fire on purpose, several times a day — comes ninth of ten. You cook for 43 minutes a day and it almost never touches you. The trampoline, the thing every parent is warned about, comes last, because almost nobody owns one.
The staircase is not close
The staircase runs at 46.28 ER visits per million hours. The ladder — an object you are visibly frightened of, that you hold on to, that you ask someone to steady — runs at 6.93.
That is a factor of 6.7. And the stairs beat the other four objects in the video combined: 46.28 against 13.33.
| object | the activity you are doing | hours/day | per million hours |
|---|---|---|---|
| Stove / oven | food preparation and cleanup | 0.71 | 0.34 |
| Kitchen knife | food preparation and cleanup | 0.71 | 2.33 |
| Bathtub / shower | grooming | 0.67 | 3.73 |
| Ladder | interior + exterior maintenance | 0.11 | 6.93 |
| Stairs | (modelled) time on stairs | 0.10 | 46.28 |
The bathtub is the one that should bother you most after the stairs. A quarter of a million visits, concentrated into the few minutes a day you are actually standing in one.
Why the floor is not on this list
Build a leaderboard by counting objects mentioned in ER records and it fills up with floors, tables and counters. Those are not hazards. They are where you land.
You can get the distinction out of the records themselves, because the word in front of the object gives it away:
| narrative | reading |
|---|---|
| fell down the stairs | the stairs supplied the fall — cause |
| fell off the ladder | the ladder supplied the height — cause |
| slipped in the shower | the tub supplied the slip — cause |
| fell on the floor | the floor is where you landed — place |
Read that preposition across every 2024 home narrative and the corpus sorts the objects for you, with no editorial judgement required:
| object | cause | place | ratio | top prepositions |
|---|---|---|---|---|
| Bathtub | 84% | 16% | 5.3× | in 46%, of 19% |
| Stairs | 75% | 25% | 3.1× | down 27%, on 11% |
| Ladder | 69% | 31% | 2.2× | off 27%, on 19% |
| Bed | 66% | 34% | 1.9× | of 53%, from 10% |
| Sofa | 54% | 46% | 1.2× | on 27%, from 20% |
| Rug / carpet | 27% | 73% | 0.4× | on 42%, over 17% |
| Door | 25% | 75% | 0.3× | on 19% |
| Table | 11% | 89% | 0.1× | on 23%, coffee 15% |
| Floor | 9% | 91% | 0.1× | to 32%, on 31% |
| Counter | 9% | 91% | 0.1× | on 43% |
Floors, counters, tables, doors and rugs are out. Note that the bed and the sofa pass this test — they are genuinely involved in how people get hurt. They are still in the data file. They just lose badly once you divide by the nine hours and the two and a half hours.
The number that could break this
One figure in this analysis is not measured. No federal survey records how long you spend on stairs. The American Time Use Survey has a line for grooming and a line for lawn care and nothing for stair traversal.
So it is modelled, at six minutes a day — and modelled generously, well above any published estimate. That choice is deliberate and it runs against the finding: a bigger denominator makes the stairs number smaller. If stairs still win under the assumption most favourable to them, the result is not an artefact of the assumption.
For the staircase to lose its top spot, the average American aged 15 or over would have to spend more than 40 minutes a day on stairs. Every day.
That is the number to argue with. Everything else in the file is published by a federal agency.
Frequency is not severity
The ranking above is how often an object hurts you. How badly is a different ordering, and the file carries both.
| object | rank by rate | severe share |
|---|---|---|
| Bed | 7 | 32.3% |
| Sofa / couch | 8 | 30.1% |
| Bathtub / shower | 3 | 23.6% |
| Ladder | 2 | 22.2% |
| Stairs | 1 | 18.9% |
| Kitchen knife | 6 | 1.6% |
The knife is sixth by rate and dead last by severity — 1.6% of knife visits end in admission. It cuts a lot of people and it almost never puts anyone in a hospital bed. The bed is seventh by rate and first by severity, because the people getting hurt falling out of one are mostly old.
The video ranks on frequency. The file lets you rank on severity and get a different answer, which is the point of shipping it.
What this does not say
NEISS counts emergency-room visits. It does not count injuries treated at urgent care, at a GP, or with a bag of frozen peas, and it does not count deaths that never reached an ER. It is a sample of about a hundred hospitals weighted up to national estimates, so every count here carries a sampling interval that this file does not propagate.
The denominators are activities, not object contact. “Hours of grooming” includes plenty of time nowhere near a bathtub. That inflates every denominator and so understates every rate — roughly uniformly, which is why the ranking survives it, but the absolute numbers should be read as floors rather than point estimates.
This is adults only, age 15 and over, matched to the time-use survey's universe. Child injuries are a different distribution and are left out rather than blended in. Home injuries only — a ladder fall at work is not in here. And the injuries are 2024 while the hours are 2025; activity budgets move slowly, but they are not the same year.
Ten objects, both rankings, the severity split and the full method — including the one number that could overturn this.
Sources: CPSC NEISS 2024 public case file, location = home, age 15+ · BLS American Time Use Survey 2025, Table A-1 · US civilian noninstitutional population 15+, 272,600,000. Every figure reproduced from the raw files by data/_neiss/rate.py, preposition.py and severity.py.