In urban crossroad tests, sirens were effective at only 26 to 40 feet, and the old mechanical wail (40 feet) barely beat the electronic wail (39 feet) (NHTSA, Lights and Siren Use by Emergency Medical Services 2017, citing Skeiber 1978). That is why a siren works as a request for attention rather than a guarantee of it. 225 people died in U.S. crashes involving emergency vehicles in 2024 (National Safety Council, Injury Facts, 2024 data), and ambulance transports with lights and sirens crashed at 17.1 per 100,000 versus 7.0 without (Watanabe et al., Annals of Emergency Medicine 2019). We aggregated data from NHTSA, the National Safety Council, the Federal Highway Administration, the National Bureau of Standards (today NIST), the U.S. Fire Administration, NIDCD, the California Code of Regulations and peer-reviewed work from the Acoustical Society of America, Annals of Emergency Medicine and Sensors. This roundup covers how far a warning reaches; for what the same sirens do to the crews riding behind them, see our firefighter and EMS siren noise statistics.
TL;DR
- In urban tests, sirens were effective at 81 to 146 feet on a straight road but only 26 to 40 feet at a crossroad (NHTSA 2017, citing Skeiber 1978).
- Electronic sirens measured 122 to 146 dB at 2 m versus 111 to 124 dB for electromechanical sirens (National Bureau of Standards 1974).
- California requires at least 120 dB(A) on axis at 3.0 m for a Class A siren (California Code of Regulations, Title 13, Section 1028).
- Every 2 dB increase in in-car music raised the siren detection threshold by 1.0 to 1.7 dB (Baba and Ebata, Acoustical Society of America 1996).
- 225 people died in crashes involving emergency vehicles in 2024 (National Safety Council, Injury Facts, 2024 data).
- 45.7% of fatal ambulance crashes in 2012 to 2018 happened in emergency use (NHTSA, DOT HS 813 480, 2023).
- More than 25% of emergency vehicle crashes occur at signalized intersections (FHWA and NHTSA, Traffic Signal Preemption 2006).
- Lights and siren are used on 74% of EMS responses and save 42 seconds to 3.8 minutes on average (Joint Statement on Lights and Siren Vehicle Operations 2022).
- Ambulance transports with lights and sirens crashed at 17.1 per 100,000 versus 7.0 without (Watanabe et al., Annals of Emergency Medicine 2019).
- 55% of adults aged 75 and older have disabling hearing loss (NIDCD, NHANES 2015-2020).
- 4.5% of drivers were visibly manipulating a handheld device at a typical daylight moment in 2024 (NHTSA, DOT HS 813 788, 2026).
- An AI system detected emergency vehicle sounds with 97.5% accuracy in real-world measurements (Banchero et al., Sensors 2025).
1. Siren Output: Electronic vs Electromechanical Sirens
Raw output was never the weak point of the electronic siren. Electronic sirens measured 122 to 146 dB at 2 m in an anechoic chamber, against 111 to 124 dB for electromechanical (motor-driven, spinning-rotor) sirens, so the loudest electronic units beat the loudest mechanical ones by 22 dB at the source (146 minus 124). On paper that headroom survives distance: the authors’ spreading adjustment to 64 m keeps the same gap, 92 to 116 dB for electronic units against 81 to 94 dB for electromechanical ones.
The regulatory floor is far below what the best hardware can do. California asks for 120 dB(A) on axis at 3.0 m for a Class A siren and lets the level fall as the listener moves off to the side, which is exactly where cross traffic sits (California Code of Regulations, Title 13, Section 1028). The rule also fixes the two familiar patterns: the slow “wail” (a long rise and fall in pitch) and the fast “yelp”.
| Metric | Value | Source |
|---|---|---|
| Mean sound level at 2 m, electronic sirens vs electromechanical sirens | 122 to 146 dB vs 111 to 124 dB | National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974 |
| Calculated level adjusted to 64 m, electronic sirens vs electromechanical sirens | 92 to 116 dB vs 81 to 94 dB | National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974 |
| Minimum level on the siren axis at 3.0 m, Class A vs Class B siren | 120 dB(A) vs 115 dB(A) | California Code of Regulations, Title 13, Section 1028 |
| Minimum level 50 degrees off the siren axis at 3.0 m, Class A vs Class B siren | 113 dB(A) vs 108 dB(A) | California Code of Regulations, Title 13, Section 1028 |
| Allowed drop during the lowest part of a wail cycle | No more than 10 dB(A) below the table values | California Code of Regulations, Title 13, Section 1028 |
| Required wail rate vs required yelp rate | 10 to 30 vs 150 to 250 oscillations per minute | California Code of Regulations, Title 13, Section 1028 |
| Octave band where the maximum on-axis level must fall | 1,000 Hz or 2,000 Hz band | California Code of Regulations, Title 13, Section 1028 |
Most recent available data: the head-to-head electronic vs electromechanical measurements come from a 1974 test of four electronic and nine electromechanical sirens. The California rule carries effective dates from 1980 to 1982. Siren output above the legal minimum is a manufacturer choice, not a requirement.
2. Vehicle Cabin Attenuation and In-Car Masking
The car body is the siren’s biggest enemy. “Insertion loss” is the acoustic term for how much sound a barrier removes, here the closed car shell between the street and the driver’s ears. Across 23 automobiles, the closed car body removed 21 to 35 dB of sound in the 1,000 Hz band, the band where siren energy peaks. A siren that loses 35 dB on the way into the cab has to compete with engine, tire and radio noise on far more even terms.
The same 1974 tests worked backward from that loss: to sound as loud inside the car as ordinary speech, a siren had to deliver 100 to 108 dB at 1,000 Hz outside the vehicle (National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974). Masking then does the rest. Japanese researchers found the detection threshold climbs 1.0 to 1.7 dB for every 2 dB of extra music, and that simply driving at medium speed, with attention on the road, pushes it up further (Baba and Ebata, Acoustical Society of America 132nd Meeting 1996). Motor officers face the reverse problem, with wind noise building up inside the helmet, as covered in our police motorcycle helmet noise statistics.
| Metric | Value | Source |
|---|---|---|
| Sound removed by the closed car body at 1,000 Hz, range across 23 automobiles | 21 to 35 dB | National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974 |
| Siren level needed outside the car at 1,000 Hz to sound as loud as ordinary speech inside | 100 to 108 dB | National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974 |
| Rise in the siren detection threshold for every 2 dB increase in in-car music level | 1.0 to 1.7 dB | Baba and Ebata, Acoustical Society of America 1996 |
| Rise in the detectable siren level in a car moving at medium speed vs a car standing still | About 0.5 dB | Baba and Ebata, Acoustical Society of America 1996 |
| Distance at which a parked driver heard an ambulance approaching at 25 mph: quiet cab (30 dB, windows open) vs noisy cab (70 dB, windows closed, loud radio, phone call) | 1,500 feet vs 10 to 283 feet | NHTSA 2017, citing MONOC 2012 |
Context note: the MONOC test was reported to NHTSA as a personal communication and the report itself calls it “less scientific” than older work. Its value is directional: cabin noise rising from 30 to 70 dB cut hearing distance from 1,500 feet to at most 283 feet. Wail and yelp got through distractions best, while hi-lo and electronic air horn did worst.
3. Audibility Distance by Siren Type and Setting
Distance is where the mechanical siren’s reputation comes from, and the data partly back it. On an urban straight road, the mechanical wail was effective at 146 feet versus 123 feet for the electronic wail, 120 for the yelp and 81 for the hi-lo (two alternating tones). The edge disappears where it matters most: at a crossroad, the four sound types landed within 14 feet of each other, all at 40 feet or less.
Suburban streets gave sirens their longest range, past 400 feet for every type except hi-lo. The federal study behind this era of data concluded that relying on audible warning devices to alert drivers in traffic was “not justified”, and that making them loud enough would create unacceptable community noise (Potter et al., Effectiveness of Audible Warning Devices on Emergency Vehicles, U.S. DOT 1977). The newest lever is low frequency: adding a Rumbler tone to a wail made jurors in a driving simulator markedly better at telling where the vehicle was coming from (Angione et al., Perceived Effectiveness of the Rumbler Emergency Siren System, INTER-NOISE 2018).
| Metric | Value | Source |
|---|---|---|
| Effective siren distance, urban straight road (window open, no radio): mechanical wail vs electronic wail vs yelp vs hi-lo | 146 vs 123 vs 120 vs 81 feet | NHTSA 2017, citing Skeiber 1978 |
| Effective siren distance, urban crossroad (window open, no radio): mechanical wail vs electronic wail vs yelp vs hi-lo | 40 vs 39 vs 38 vs 26 feet | NHTSA 2017, citing Skeiber 1978 |
| Effective siren distance, suburban straight road (window open, no radio): mechanical wail vs electronic wail vs yelp vs hi-lo | 445 vs 440 vs 426 vs 257 feet | NHTSA 2017, citing Skeiber 1978 |
| Loss from ground reflection at certain siren-to-listener distances | Up to 60 dB | National Bureau of Standards, Summary Report on Emergency Vehicle Sirens 1974 |
| Gain in correct judgments of the emergency vehicle’s direction when a low-frequency Rumbler tone is added to the wail (23 jurors) | Nearly 25% higher | Angione et al., INTER-NOISE 2018 |
Most recent available data: the siren-type distance tables date from a 1977 study published for the U.S. DOT in 1978, and no comparable large field study has been published since. The ground-reflection figure describes sharp dips at specific distances, where the direct and road-reflected sound cancel; a sweeping siren pitch makes those dips shallower in practice. The Rumbler result comes from a simulator, not real intersections.
4. Intersections: Where Warning Distance Runs Out
Put sections 2 and 3 together and the intersection problem is arithmetic: a driver behind closed windows, with cross traffic approaching from the side, gets the shortest warning distance in the whole system. More than 25% of all emergency vehicle crashes occur at signalized intersections, according to the federal cross-cutting study of signal preemption (FHWA and NHTSA, Traffic Signal Preemption for Emergency Vehicles 2006).
Preemption, the technology that turns the light green for the approaching emergency vehicle instead of asking the siren to stop cross traffic, has the strongest before-and-after numbers in this roundup. St. Paul and Plano both equipped every signal, and both saw emergency vehicle crashes fall sharply; a 1978 Denver study cited in the same report found response times dropped 14 to 23 percent. The crash pattern is consistent with that diagnosis: in a survey of 675 EMS practitioners, 80% of those involved in a crash reported a broadside strike (a side impact, the classic intersection collision).
| Metric | Value | Source |
|---|---|---|
| Share of emergency vehicle crashes that occur at signalized intersections | More than 25% | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| Emergency vehicle crashes involving fatalities per year in the U.S. | Approximately 80 | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| St. Paul emergency vehicle crashes: 1967 high vs average after preemption spread to 285 of 308 intersections | 8 vs 3.3 per year | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| Plano emergency vehicle crashes at intersections, before vs after preemption | 2.3 per year vs fewer than one every five years | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| Traffic signals equipped with preemption nationwide | Over 30,964 in 375 jurisdictions | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| Fatal-crash investigations (27 cases) citing improper clearing of an intersection vs proceeding through a red light | 14.8% vs 7.4% | NHTSA, DOT HS 813 480, 2023 |
| EMS practitioners in a crash who reported a broadside strike | 80% | Joint Statement on Lights and Siren Vehicle Operations 2022 |
Most recent available data: the preemption figures come from a 2006 federal study, and about 20 percent of traffic signals in the 78 largest metropolitan areas had preemption at the time. Treat the St. Paul and Plano numbers as case studies, not national effects.
5. Emergency Vehicle Crash and Fatality Toll
The cost of a missed warning falls mostly on people outside the emergency vehicle. Of the 225 people killed in U.S. crashes involving emergency vehicles in 2024, 56% were occupants of other vehicles and 24% were pedestrians, while emergency vehicle drivers were 9% and 70% of all deaths came in multi-vehicle crashes (National Safety Council, Injury Facts: Emergency Vehicles, 2024 data). Crashes involving police vehicles accounted for two thirds of those deaths (150 of 225).
Ambulance-specific federal data say the same thing from a different angle: other-vehicle occupants were 52.3% of the dead in 2012 to 2018, and almost half of fatal crashes happened in emergency use (NHTSA, Analysis of Ground Ambulance Crash Data From 2012 to 2018, DOT HS 813 480, 2023). The emergency-use share was higher in the older 1992 to 2011 series (NHTSA, Ground Ambulance Crashes 2014). For responders themselves, vehicle collisions were the third leading cause of on-duty firefighter deaths in 2024 (USFA, Firefighter Fatalities in the United States in 2024).
| Metric | Value | Source |
|---|---|---|
| Deaths in U.S. crashes involving emergency vehicles, 2024 | 225 | National Safety Council, Injury Facts, 2024 data |
| 2024 deaths in crashes involving police vehicles vs ambulances vs fire trucks | 150 vs 47 vs 28 | National Safety Council, Injury Facts, 2024 data |
| Fatal ambulance-involved crashes per year vs fatalities per year, 2012 to 2018 average | 24.7 vs 28.4 | NHTSA, DOT HS 813 480, 2023 |
| Fatal ambulance crashes with the ambulance in emergency operation (2012 to 2018) vs with lights and sirens active (2013 to 2018) | 45.7% vs 28% | NHTSA, DOT HS 813 480, 2023 |
| Ambulance injury crashes with lights and sirens active, 2013 to 2018 estimate | About 36.2% | NHTSA, DOT HS 813 480, 2023 |
| Ambulance crashes per year, 1992 to 2011 average | About 4,500 | NHTSA, Ground Ambulance Crashes 2014 |
| Fatal ambulance crashes in emergency use, 1992 to 2011 | 58% | NHTSA, Ground Ambulance Crashes 2014 |
| Firefighters killed in vehicle collisions, 2024 | 8 of 72 on-duty deaths (11%) | USFA, Firefighter Fatalities in the United States in 2024 |
Context note: USFA adds that almost one quarter (24%) of firefighters killed in vehicle collisions from 2015 to 2024 were in a privately owned vehicle, which usually carries no siren at all.
6. Lights and Siren Use: Minutes Saved vs Crash Risk
If the siren buys so little warning distance, the question becomes what it buys in time. Lights and siren save an average of 42 seconds to 3.8 minutes per response or transport across more than a dozen studies, yet they are used on 74% of EMS responses (Joint Statement on Lights and Siren Vehicle Operations on EMS Responses, NAEMSP and partner organizations, 2022). Only 6.9% of those responses end in a potentially lifesaving intervention, which is why 14 EMS and fire organizations now argue for using them selectively.
The crash side of the ledger comes from 2016 national EMS records covering about 19 million 911 scene responses. The adjusted odds of a crash rise by half during responses and nearly triple during transports (Watanabe et al., Annals of Emergency Medicine 2019). The “wake effect” adds hidden cost: crashes between other drivers reacting to the emergency vehicle, which never appear in emergency vehicle crash counts.
| Metric | Value | Source |
|---|---|---|
| EMS responses vs EMS transports that use lights and siren | 74% vs 21.6% | Joint Statement on Lights and Siren Vehicle Operations 2022 |
| Average time saved with lights and siren, range across studies | 42 seconds to 3.8 minutes | Joint Statement on Lights and Siren Vehicle Operations 2022 |
| Lights and siren responses that led to a potentially lifesaving intervention | 6.9% | Joint Statement on Lights and Siren Vehicle Operations 2022 |
| Ambulance crash rate per 100,000 responses, without vs with lights and sirens | 4.6 vs 5.4 (adjusted OR 1.5; 95% CI 1.2 to 1.9) | Watanabe et al., Annals of Emergency Medicine 2019 |
| Ambulance crash rate per 100,000 transports, without vs with lights and sirens | 7.0 vs 17.1 (adjusted OR 2.9; 95% CI 2.2 to 3.9) | Watanabe et al., Annals of Emergency Medicine 2019 |
| ”Wake effect” collisions among other vehicles for every crash involving an emergency vehicle | Approximately four | Joint Statement on Lights and Siren Vehicle Operations 2022 |
| Lights and siren use after some agencies cut it back, with no discernible harm to patient outcomes | 20 to 33% | Joint Statement on Lights and Siren Vehicle Operations 2022 |
Context note: Watanabe et al. is observational, so it shows association rather than cause, and the adjusted figures are odds ratios rather than simple multiples of risk.
7. The Listener Side: Hearing, Distraction, and In-Car Siren Detection
Every figure above assumes a listener with normal hearing who is paying attention. Neither assumption holds for a large share of drivers. 55% of U.S. adults aged 75 and older have disabling hearing loss, defined as 35 dB or more in the better ear (NIDCD, Quick Statistics About Hearing), so the 21 to 35 dB lost through the car body (see section 2) lands on top of an already raised threshold. Our presbycusis statistics track how that threshold climbs with age.
Distraction is growing in national roadside observations: drivers seen manipulating a handheld device jumped from 3.0% in 2023 to 4.5% in 2024 (NHTSA, Driver Electronic Device Use in 2024, DOT HS 813 788). The engineering answer is to have the car listen for the driver. A 2025 system using four microphones detected emergency vehicle sounds at 97.5% accuracy in real-world measurements (Banchero et al., Sensors 2025), but microphones inside the cabin hear the same muffled siren the driver does (Cantarini et al., Sensors 2022). Quieter electric cars face the mirror problem, being too hard to hear, covered in our EV acoustic warning statistics.
| Metric | Value | Source |
|---|---|---|
| U.S. adults with disabling hearing loss, ages 65 to 74 vs 75 and older | 22% vs 55% | NIDCD, NHANES 2015-2020 |
| U.S. adults reporting some trouble hearing | About 15% (37.5 million) | NIDCD, National Health Interview Survey 2012 data |
| Drivers visibly manipulating a handheld device at a typical daylight moment, 2023 vs 2024 | 3.0% vs 4.5% | NHTSA, DOT HS 813 788, 2026 |
| Drivers aged 16 to 24 visibly manipulating a handheld device, 2024 | 8.7% | NHTSA, DOT HS 813 788, 2026 |
| Drivers speaking with a visible headset, 2024 | 0.5% | NHTSA, DOT HS 813 788, 2026 |
| AI detection accuracy for emergency vehicle sounds, simulated data vs real-world measurements | 98.86% vs 97.5% | Banchero et al., Sensors 2025 |
| AI direction error when locating the emergency vehicle, simulations vs real world | 5.12 vs 10.30 degrees | Banchero et al., Sensors 2025 |
| Siren detection score (AUPRC, from 0 to 1) with the microphone inside the cabin vs behind the license plate | 0.73 vs 0.86 | Cantarini et al., Sensors 2022 |
Context note: AUPRC (area under the precision-recall curve) is a detection score where 1 is perfect. The NIDCD trouble-hearing figure is self-reported and dates from 2012; the disabling-loss figures are from audiometric exams.
Summary: Emergency Vehicle Siren Audibility by the Numbers
| Metric | Value | Source |
|---|---|---|
| Effective siren distance at an urban crossroad, range across four siren sounds | 26 to 40 feet | NHTSA 2017, citing Skeiber 1978 |
| Sound removed by the closed car body at 1,000 Hz, across 23 automobiles | 21 to 35 dB | National Bureau of Standards 1974 |
| Minimum Class A siren level on axis at 3.0 m | 120 dB(A) | California Code of Regulations, Title 13, Section 1028 |
| Rise in siren detection threshold per 2 dB of in-car music | 1.0 to 1.7 dB | Baba and Ebata, Acoustical Society of America 1996 |
| Gain in correct direction judgments with a Rumbler tone added to the wail | Nearly 25% higher | Angione et al., INTER-NOISE 2018 |
| Deaths in U.S. crashes involving emergency vehicles, 2024 | 225 | National Safety Council, Injury Facts, 2024 data |
| 2024 deaths: police vehicles vs ambulances vs fire trucks | 150 vs 47 vs 28 | National Safety Council, Injury Facts, 2024 data |
| Fatalities per year in ambulance crashes, 2012 to 2018 average | 28.4 | NHTSA, DOT HS 813 480, 2023 |
| Fatal ambulance crashes in emergency operation, 2012 to 2018 | 45.7% | NHTSA, DOT HS 813 480, 2023 |
| Emergency vehicle crashes at signalized intersections | More than 25% | FHWA and NHTSA, Traffic Signal Preemption 2006 |
| Average time saved with lights and siren | 42 seconds to 3.8 minutes | Joint Statement on Lights and Siren Vehicle Operations 2022 |
| Ambulance crash rate per 100,000 transports, without vs with lights and sirens | 7.0 vs 17.1 | Watanabe et al., Annals of Emergency Medicine 2019 |
| Firefighters killed in vehicle collisions, 2024 | 8 of 72 (11%) | USFA, Firefighter Fatalities in the United States in 2024 |
| U.S. adults aged 75 and older with disabling hearing loss | 55% | NIDCD, NHANES 2015-2020 |
| Drivers visibly manipulating a handheld device, 2024 | 4.5% | NHTSA, DOT HS 813 788, 2026 |
| AI detection accuracy for emergency vehicle sounds, real-world measurements | 97.5% | Banchero et al., Sensors 2025 |
| Siren detection score, microphone inside the cabin vs behind the license plate | 0.73 vs 0.86 | Cantarini et al., Sensors 2022 |
Methodology and Sources
- NHTSA Office of EMS, Lights and Siren Use by Emergency Medical Services (EMS): Above All Do No Harm (Kupas), May 2017: siren effectiveness distances (as reported in Skeiber 1978) and the MONOC 2012 hearing-distance test.
- National Bureau of Standards (Jones, Quindry and Rinkinen), Summary Report on Emergency Vehicle Sirens, LESP-RPT-0502.00, 1974: electronic vs electromechanical output, car insertion loss, ground reflection.
- Potter, Fidell, Myles and Keast, Effectiveness of Audible Warning Devices on Emergency Vehicles, U.S. DOT, DOT-TSC-OST-77-38, 1977
- California Code of Regulations, Title 13, Section 1028, Siren Performance Requirements
- Baba and Ebata, Study on detection of the electronic siren from an ambulance, Acoustical Society of America 132nd Meeting, 1996
- Angione et al., Perceived Effectiveness of the Rumbler Emergency Siren System, INTER-NOISE 2018
- National Safety Council, Injury Facts: Emergency Vehicles (analysis of NHTSA FARS, 2024 data)
- NHTSA, Analysis of Ground Ambulance Crash Data From 2012 to 2018, DOT HS 813 480, July 2023
- NHTSA, The National Highway Traffic Safety Administration and Ground Ambulance Crashes, April 2014
- U.S. Fire Administration and National Fallen Firefighters Foundation, Firefighter Fatalities in the United States in 2024
- FHWA and NHTSA, Traffic Signal Preemption for Emergency Vehicles: A Cross-Cutting Study, FHWA-JPO-05-010, January 2006
- Joint Statement on Lights and Siren Vehicle Operations on EMS Responses, NAEMSP and partner organizations, February 2022
- Watanabe et al., Is Use of Warning Lights and Sirens Associated With Increased Risk of Ambulance Crashes? Annals of Emergency Medicine 74(1), 2019
- NIDCD, Quick Statistics About Hearing, Balance, and Dizziness
- NHTSA, Driver Electronic Device Use in 2024, DOT HS 813 788, April 2026
- Banchero, Vacalebri-Lloret, Mossi and Lopez, Enhancing Road Safety with AI-Powered System for Effective Detection and Localization of Emergency Vehicles by Sound, Sensors 25(3), 2025
- Cantarini, Gabrielli and Squartini, Few-Shot Emergency Siren Detection, Sensors 22(12), 2022
- Data watch: the core acoustic evidence (siren output, car insertion loss, distance by siren type) is from 1974 to 1978 and has not been repeated at scale with modern, better-insulated cars, so today’s real attenuation may be higher than the 21 to 35 dB measured then. The National Safety Council page quotes the Watanabe lights-and-siren crash rates as 5.5 and 16.5 per 100,000, while the published paper reports 5.4 and 17.1; we use the paper. NHTSA’s 2017 report lists lights-and-siren transport use as 22.7% in one table and 21.6% in another for 2015 NEMSIS data; we quote the joint statement’s 21.6%. The FHWA 2006 report contains a garbled siren loudness figure (“1,200 db”), which we did not use. Banchero et al. give 97.5% as real-world accuracy in the abstract but describe the same figure differently in the results section. NHTSA’s ambulance averages (28.4 deaths per year, 2012 to 2018) and the NSC count (47 deaths in 2024) use different years and case definitions and should not be read as a trend. A national FARS update for 2025 is expected.
Last updated: October 24, 2026. We update this roundup quarterly, and the next refresh is expected when NHTSA releases the 2025 FARS Annual Report File and the National Safety Council updates its Injury Facts emergency vehicle page.